Darwin in the Genome Molecular Strategies in Biological Evolution
Lynn Helena Caporale
McGraw-Hill New York Chicago San Francisco Lisbon London Madrid Mexico City Milan New Delhi San Juan Seoul Singapore Sydney Toronto
Copyright © 2003 by Lynn Helena Caporale. All rights reserved. Manufactured in the United States of America. Except as permitted under the United States Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or by any means, or stored in a database or retrieval system, without the prior written permission of the publisher. 0-07-142051-7 The material in this eBook also appears in the print version of this title: 0-07-137822-7.
All trademarks are trademarks of their respective owners. Rather than put a trademark symbol after every occurrence of a trademarked name, we use names in an editorial fashion only, and to the benefit of the trademark owner, with no intention of infringement of the trademark. Where such designations appear in this book, they have been printed with initial caps. McGraw-Hill eBooks are available at special quantity discounts to use as premiums and sales promotions, or for use in corporate training programs. For more information, please contact George Hoare, Special Sales, at
[email protected] or (212) 9044069.
TERMS OF USE This is a copyrighted work and The McGraw-Hill Companies, Inc. (“McGraw-Hill”) and its licensors reserve all rights in and to the work. Use of this work is subject to these terms. Except as permitted under the Copyright Act of 1976 and the right to store and retrieve one copy of the work, you may not decompile, disassemble, reverse engineer, reproduce, modify, create derivative works based upon, transmit, distribute, disseminate, sell, publish or sublicense the work or any part of it without McGraw-Hill’s prior consent. You may use the work for your own noncommercial and personal use; any other use of the work is strictly prohibited. Your right to use the work may be terminated if you fail to comply with these terms. THE WORK IS PROVIDED “AS IS”. McGRAW-HILL AND ITS LICENSORS MAKE NO GUARANTEES OR WARRANTIES AS TO THE ACCURACY, ADEQUACY OR COMPLETENESS OF OR RESULTS TO BE OBTAINED FROM USING THE WORK, INCLUDING ANY INFORMATION THAT CAN BE ACCESSED THROUGH THE WORK VIA HYPERLINK OR OTHERWISE, AND EXPRESSLY DISCLAIM ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. McGraw-Hill and its licensors do not warrant or guarantee that the functions contained in the work will meet your requirements or that its operation will be uninterrupted or error free. Neither McGraw-Hill nor its licensors shall be liable to you or anyone else for any inaccuracy, error or omission, regardless of cause, in the work or for any damages resulting therefrom. McGraw-Hill has no responsibility for the content of any information accessed through the work. Under no circumstances shall McGraw-Hill and/or its licensors be liable for any indirect, incidental, special, punitive, consequential or similar damages that result from the use of or inability to use the work, even if any of them has been advised of the possibility of such damages. This limitation of liability shall apply to any claim or cause whatsoever whether such claim or cause arises in contract, tort or otherwise. DOI: 10.1036/0071420517
Want to learn more? We hope you enjoy this McGraw-Hill eBook! , If you, d like more information about this book, its author, or related books and websites, please click here.
With love, to Mom and Dad and to our next generation— Brooks, Michael, Parker, and Rockella
This page intentionally left blank.
For more information about this title, click here.
Contents
Preface
ix
Prologue
Chance Favors the Prepared Genome
Chapter 1
Diversity or Death
Chapter 2
The Magic Staircase
Chapter 3
Predators Battle Prey in the Genome
Chapter 4
Mutation Is Not Monotonous
Chapter 5
What the Antibody Genes Tell Us
Chapter 6
Slippery DNA and Tuning Knobs
Chapter 7
Everyone Has Something to Teach Us
Chapter 8
The Genome Sends an SOS
Chapter 9
Journeys through Space and Time
Chapter 10
Strategies as Targets, Round One: The Pathogens
Chapter 11
Strategies as Targets, Round Two: Cancer
9 19
Theme and Variations
Chapter 13
Family Heirlooms: A Framework for Evolution Interchangeable Parts
27
35 45 59 71
85 97
103
109
Chapter 12
Chapter 14
1
119 125
133
v
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
vi
CONTENTS
Chapter 15
Jumping Genes
145
Chapter 16
Be Prepared for the Unexpected
Chapter 17
Mixing Up Genes for the Children
Chapter 18
Sculpting the Genome
Chapter 19
Decoding DNA
187
What Became of the First Genome?
Glossary
203 209
Index
237
Acknowledgments
243
165
175
Epilogue
Notes
155
195
Preface I was in a new world, and . . . could not help speculating on what my wanderings there . . . might bring to light. —Alfred Russel Wallace1
L
ong ago, consciousness began to emerge within life on Earth. Now, at the very moment when we are alive, consciousness is starting to comprehend the extraordinary information that each of us carries inside on the thin ribbons of our genome. What will we find, now that we can peer so deep within ourselves? Just four hundred years ago, lenses were ground, placed together in a tube, and turned toward the night sky. Galileo’s revelations did not require the see-to-the-edge-of-time spacecraft-mounted telescopes of our day. He looked through lenses with the strength of the 12× binoculars that today we carry so casually to examine the markings on a bird overhead. By turning such simple lenses to the sky, he showed us something that directly contradicted the overwhelming evidence of our senses. Contrary to what seemed so obvious every day from the sunrise, to the movement of the sun across the sky, to its setting on the opposite horizon, and its predictable rise the next day, we learned that the sun does not revolve around the Earth. We were certain that the sun circled around us, but in fact it does not. When we learned that we are not at the very center of all that there is, our view of who we are and our place in the universe was dramatically changed (although there was, and in many ways still is, resistance to this change in perspective). Yet this revolution in our comprehension of the world and our place in it was, I believe, small, very small, compared to the change in vii
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
viii
PREFACE
our comprehension of the world and our place in it that is about to occur as we look inside at our own genome, and at the genomes of other creatures. The initial impulse for this book came when I organized and cochaired a conference that may well represent a landmark in discussions of evolution. This gathering was filled with enthusiastic discussions among people whose work varies from sitting at computers to studying the life that lives within beautiful shells along the beach. I was convinced that the research we discussed, and, since then, the new discoveries that continue to capture our imagination, could be shared with a broad range of readers. For nonscientists, I worked hard to break through the barrier of technical language to share the heart of these ideas and discoveries with you, but if you find that I have failed to do this in some spots, I also worked hard to make it possible to skip some sections and pick up the train of thought. There is a glossary in the back to help out, too. For readers who are professional scientists, I ask of course for patience as I explain some things that you may already have learned. (For everyone, I have inserted a few brief word games, which a friend of mine dislikes intensely and has characterized as “typos on steroids”; if you agree with him, jump over them.) I ask all readers to share with me in reflecting on what the implications of these new discoveries might be. On the shelves of a bookstore, imaginative writings are thought to be confined to the fiction section, but in fact imagination drives scientists as much as it does novelists. Indeed, much as a novel is enhanced when its locations are described in factual detail and its characters made to seem real, so nonfiction can be more inspiring when the reader is taken into the imagination of the scientist. In such writing, as in the mind of a good scientist, the boundaries between data and imagination must be made very clear. In this book, I have worked to weave a wide range of research into a coherent vision, while describing much that remains the subject of active current investigation and even controversy. Some sections, which it is my job to indicate clearly, step into the realm of scientists’ imagination, suggesting where new discoveries may lead. As one of the founders of the field of molecular biology, Max Delbrück, pointed out, “Any living cell carries within it the results of billions of years of experimentation by its ancestors.” The steps of this long journey are inscribed in our DNA, and in the DNA of all the living creatures that share our time on Earth. There are messages in our genome for each of us; they come from our diverse ancestors, and they outline our path across past ages, in other forms, to here. If we strain to listen, will our genome help us to comprehend how mutation and selection ever could have led to us—to
PREFACE
ix
me, sitting here and writing, and to you, sitting there and reading? I have written this book to explore these very questions. Our genome encodes many types of information. Perhaps the best studied is the information that encodes our proteins. Other information controls which proteins are made in a cell in our heart, or in our brain, or in our thumbs. But it is clear that there is room in our genome for other types of information that we are only now beginning to perceive. This information can be found in the spaces between genes, and even doubled up along the stretches of DNA that encode our proteins. Early examples of such multiple messages involved information encoding proteins that ran together along a piece of DNA in different reading frames, something like you find tthhee itrw ol emtetsesrasg easl tmeirxneadt iunpg here. In fact, it was this ability to transmit more than one message through the same stretch of DNA that first led to my writing about evolution, as described in Chapter 13.2 As will be explained in this book, I have proposed that this extra coding potential has been harnessed3 by natural selection to improve the probability of survival. Some scientists may disagree with my interpretation of the data before us, leading to interesting discussions about data; in studies of the genome, data are coming very quickly now. Some people do seek data that show that their ideas are right, and discard data that are not consistent with their preconceived ideas, but that is not science. Because this book mentions evolution, it may attract religious debate, but faith is, of course, something entirely different. This book addresses scientific research, not faith. But knowledge does have implications. Rather than view evolution as a no-holds-barred fight to the death, I find that the discoveries described in this book harmonize with the teachings of many great religions: that people are brothers and sisters; that we should respect, welcome, and share with others; and that reverence should extend to other forms of life. Indeed the intellectual, practical, political, and philosophical implications of being able to read the information within genomes reach into everyone’s lives and emphasize the importance of diversity for survival. This includes biodiversity and, especially, should lead us to treasure the diversity of our own species. On a practical level, the knowledge that we gain from looking into our genome will facilitate medical research and also may teach us ways to design novel coding systems for our computers. By bringing our consciousness into the analysis of microbe and tumor genomes, our new skills should strengthen our hand in the age-old battle against these adversaries, as discussed in this book. Some private messages in each person’s genome may
x
PREFACE
warn of susceptibility to certain diseases, but other information will reassure people that there are other diseases that they are less likely to suffer. Many observations lead to caution in the face of proposals to fix “errors” in the human genome. When I was a little girl sitting at my desk in school, I was asked to draw a circle around a picture of a ball and to connect it, with a line, to the word ball. I assume you too had workbooks in which you connected pictures and words, one at a time. In retrospect, it was a lot like that when we first began to look at our DNA, connecting one gene to one protein. But now that we finally can see the genome as a whole, we are learning new ways to read DNA that go beyond letters and words, to understand decision-making networks and hierarchies.4 Physically, a genome may be a string of letters along a strand of the DNA double helix, but functionally it is an interconnected, highly cross-referenced system. In the more than 3 billion letters of our genome, there are concepts to pull out, ideas in one place that are connected to and built up from ideas in other places. All of biology, including our journey to this place, is waiting within us to be pieced together into an integrated living, whole. I feel fortunate to have become a biochemist at this extraordinary moment in history, and I invite you to join me on this journey of discovery into our genome, to explore what our “wanderings there might bring to light.”
Prologue Chance Favors the Prepared Genome Delicate, elusive . . . is that mysterious principle known as “Organization,” which leaves all other mysteries concerned with life stale and insignificant by comparison. —Loren Eiseley, The Immense Journey
T
here was a moment in time when the dust itself edged, in slow motion, over a boundary into life. It entered onto a path strewn with dangers, uncertainty, and creativity. It spread its growing skill across the Earth until it learned how to fly at will and how to sit still here, and to discuss its own evolution. Those who do not believe that we have evolved from life forms that are invisible to the naked eye—and even those who do—find it hard to conceive of how this journey, relying only on random mutation and survival of the “fittest,” could have succeeded. It seems almost inconceivable that there has been enough time for mere molecules to organize themselves into a being that could compose music, travel to the moon and back, and indeed analyze its own genome. How did we get this far, even once, in only billions of years? How could we have happened so randomly? In the greatest achievement in human intellectual history, the information in our genome, the product of billions of years of evolution, is now opening before us. What answers does our genome hold to the Big Questions that whisper to us all? What does it say of the immense journey it has taken as it has passed through uncounted life forms to be carried within us as we sit here, discussing our origins and our fate? 1
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
2
DARWIN IN THE GENOME
It has been well over a century since Charles Darwin and Alfred Russel Wallace first proposed that from among the natural variations in a population, the “most fit” individuals would tend to survive in greater numbers, passing their selected variations on to their progeny. But when Darwin and Wallace proposed that evolution happens through variation and then selection, they did not know the mechanism by which the varied traits were inherited. Science at that time was still was about a century away from uncovering the chemistry of heredity. And, although Darwin and Wallace had each traveled from England to the Pacific Ocean, they had not traveled to Gregor Mendel’s garden at Brno and so had never talked to their contemporary about how traits might be transmitted between generations. (What a conversation that would have been!) For his part, Mendel went beyond the general understanding that we resemble our parents and showed that traits, such as whether the peas borne by the plants in his garden were wrinkled or smooth in form or yellow or green in color, were inherited independently of each other in a way that was predictable.1 Mendel used careful statistics to work out the “laws” of heredity. He could anticipate the mix of smooth and wrinkled, yellow and green peas borne by tall and dwarf plants from generation to generation, but he could not explain how this worked. We now know that the predictability of the peas comes from the fact that the inherited variations of the peas are encoded in separately assorting stretches of the DNA of pea chromosomes. Of course, the naturalists and the statistician/monk did not have a chance to travel forward in time to kick ideas around with researchers who studied the mechanisms of mutation, but some decades after Darwin, Mendel, and Wallace were gone, their ideas were incorporated into the then more recent work to form our current understanding of evolution: Variations among organisms are due to variations in genes; variation is due to different “versions” of genes (giving, for example, green peas or yellow peas, green eyes or blue eyes) and mutation of the DNA that makes up our genes. From this variation, selection picks the most “fitted.” When we say “fit” in our daily lives, this usually leads into a conversation about workouts, but for the moment I would like to talk about personality quirks. Imagine that a developer built new homes in an isolated canyon. The homes were very desirable, but the developer had a prohibitive personality quirk: He would allow only certain cars through the gate on the road that led to the canyon. No one understood why their car was allowed through the gate or turned away, but I can tell you: The imaginary developer let families get through the gate only if at least one child in the
CHANCE FAVORS THE PREPARED GENOME
3
car was wearing a blue T-shirt. The natural consequence was that among the people who settled in the valley, there was a higher proportion of children who tended to wear blue T-shirts than there were in the general population. After a couple of generations, the isolated canyon became crowded, and the developer’s grandchildren built homes in the next canyon over. The developer’s grandchildren inherited the developer’s quirky attention to T-shirts, but they were so used to blue, from its enrichment in the first canyon, where they too lived, that they did not favor blue at the new barrier. They would admit only families that had a child wearing a yellow T-shirt. After generations of families getting through gates controlled by generations of quirky developers, the tenth canyon was inhabited by families that, without thinking about it, each day dressed their several children in different-colored T-shirts. These colorful families were the ones that had the best chance of having one of their children wearing the “right” color T-shirt and thus being able to get across barrier after barrier, from the canyon enriched in blue to the canyon enriched in yellow and beyond. As genomes travel from generation to generation across evolutionary time, we face something far more serious than quirky developers. Pathogens are among the life-and-death challenges that allow only some children to pass into the next generation. Just as the imaginary developers let families who varied the colors of their children’s T-shirts get through, pathogens let genomes get through that varied their progeny enough to avoid the pathogens’ biochemical tricks. You can be too careful. True, a genome must be conserved as it is passed from generation to generation. To reproduce a genome, it is necessary to be careful in copying DNA and to repair errors. But having all progeny be exactly the same may not be the safest strategy. In fact, Darwin and Wallace and their contemporaries were impressed with the tremendous variation they observed within each species, from birds to beetles. Now, the genome, which had been hidden from their view, is becoming a landscape for a new kind of naturalist to explore. This naturalist views the variations within each species using a different kind of binoculars. Far from being carried on a strap, these “binoculars” involve laboratory infrastructure and computers to read and analyze each species’ DNA. The variation that these naturalists study is not limited to that between feathers and limbs; rather, what captures our imagination is the variation among worm or flower or human, or bacterial genomes. What has made evolution so hard for many to accept is the assumption that it depends upon random mutation for the generation of new variations. Momentarily sloppy, the gene-copying mechanism drops something, messes
4
DARWIN IN THE GENOME
up, and passes on a mistake to a probably unfortunate member of the next generation. Through sheer luck, the change in the DNA, the accident, may turn out OK, so that the child who inherits it survives and passes it on. Rarely, through even greater luck, the change may turn out to be for the better; with those rare lucky accidents, the random mistake makes a fitter child (or fawn or tadpole or sprout or bacterium), one that is favored by natural selection. Slowly, over unfathomable lengths of time, from one rare, lucky mutation after another, these rare fitter children in turn give birth rarely and accidentally to even fitter children; and so, at its stumbling pace, evolution proceeds, selecting any advantage in a wing or a protein, one by one. That variation comes from random mutation of DNA was not, of course, Darwin’s proposal. But just as Darwin and Wallace could not incorporate into their theories what they did not know about genes, when our current understanding was developed, there was a lot that we did not know about genomes. In this book I will propose that it is time to incorporate our new discoveries into our understanding of evolution. As Baldomero Olivera, whose work will be discussed in Chapter 3, pointed out when describing his observations, “Unconventional hypotheses for these unusual data merit serious consideration.” The work described in this book has led me to the conclusion that natural selection must work not just on each individual mutation, but also on the very mechanisms that generate genetic variation—as it does on all biological functions. The research discussed in this book leads to the conclusion that mutations are not all accidents and that mutations are not always random. Our genomes, and those of other life forms, have evolved mechanisms that create different kinds of mutations in their DNA, and they reuse and adapt useful pieces of DNA, even to the point that there are genomic “interchangeable parts.” Biochemical mechanisms can arise that tend to focus genetic variation, resulting in “hot spots” of genetic change at certain places in the genome. The probability of genetic change at any given point in the genome is dependent upon the surrounding sequence of the DNA, the environment, and the proteins that are present in the cell that interact with the DNA; for example, specific types of mutation can be increased in our immune system. Evolution may not have been reaching for the goal of two eyes and a brain and two arms and two legs, but it didn’t just stumble onto us through clumsy wandering. Randomness fades in a world that rewards each step of getting better at finding food, avoiding predators, or adapting to recurring challenges. As the dust organized, it faced selection. Over time, there emerged something that, viewing the effects now, we might call strategies—
CHANCE FAVORS THE PREPARED GENOME
5
such as the ability to actively generate diversity—that enabled life to emerge from the darkness of random wandering. Because the mechanisms that change the genome fall under selective pressure, I propose, based on the new observations discussed in this book, that information can flow back from survival to the places in the genome that affect the generation of the diversity that we see around us, and that this will make genomes become more efficient at adapting and evolving. If one of the predictable characteristics of the world is that it changes over the course of generations, natural selection will lead to organisms that are more efficient at adapting to an environment that may change. These discoveries do not refute the theory of natural selection developed by Darwin and Wallace, but instead provide a deeper understanding of how natural selection leads to organisms that are better adapted to their world. Natural selection acts on all biological properties. That means that natural selection acts not only on fins and wings, but also on the mechanisms that change a genome. With time, it turns out, the “fittest” genomes, the “successful” genomes—the ones that survive—are the genomes that evolve what here I will call mutation strategies. Some readers may disagree with this use of the word strategies, as I am, after all, discussing groups of molecules. But I use this word to emphasize that the molecular mechanisms I will describe in this book have the effect of anticipating and responding to challenges and opportunities that continue to emerge in the environment. The first strategy for survival clearly is to generate diversity. The longterm survival, or fitness, of a genome often depends upon the diversity of its descendants. Genomes have evolved biochemical mechanisms that actively diversify themselves. The more diverse the progeny, the better the chance that at least some progeny will be different in a way that allows them to survive or even thrive, whether they are in an isolated canyon, a salt cave, a hot spring in Yellowstone National Park, or an irradiated can of meat—for life can survive in all of these places. Miroslav Radman, whose work is discussed in Chapter 8, described it this way: “The generation of a large repertoire of biological diversity [is] the evolutionary equivalent of buying a large number of lottery tickets.”2 The lottery winners are those who survive natural selection. You don’t want everyone in your family to buy a ticket that has the same number. We must have sunlight to construct vitamin D in our skin, but too much sun will burn us. With dark skin to protect against sunburn, a child can survive at the Equator. With lighter skin to let in more sunlight, a child can be wrapped in warmer and warmer coverings and begin a journey away from the Equator. The world inhabited by these children can be very different
6
DARWIN IN THE GENOME
from the world that best suited their parents. If the world may become different, it is an advantage for some children to be different too. If the ability to generate diversity is a useful skill, the fact that genomes can generate genetic diversity in more than one way provides an even greater advantage.3 To prepare for various levels of selection, the genome too can change just a little, a little more, or still more between generations. Maybe one quirky developer admits applicants to the new canyon not on the basis of the color of their shirts, but on the style of their top shirt buttons. Another issue in getting into some canyons may be the size of the car (too big?) or the strength of its engine (too weak?). A genome’s ability to grow and to explore new organizational structures would be severely constrained if its options were limited to changes in the molecular equivalent of the top shirt button. A single letter along a strand of the DNA double helix can change to another letter, but also a patch of letters may expand, be replaced, or be removed, or some pieces of the genome might be rearranged. By now a diverse set of biochemical mechanisms of change has emerged, each mechanism generating a different type, rate, and extent of diversity. A second strategy that has emerged in genomes, and that will be discussed starting with Chapter 7 in this book, is the reuse of useful pieces of genetic information. This is exemplified by the spread among bacteria of information that encodes resistance to antibiotics. Other useful genetic information, such as how to digest a new food source, also can come into a bacterial genome from outside—in other words, from a genome that was not its parent. As in a port city in a nation of immigrants, within bacteria new genetic ideas arrive, are put together, survive, prosper, and can thrive. Useful genetic information that is already within a genome also can be adapted for a new job, molded by making an extra copy of a piece of DNA, moving the copy around, and tinkering. Shuffling DNA around within our genome can have risks and do damage, leading, in people, to “birth defects.” But, as the roof falls, albeit ever so slowly, in, sitting still also has its risks. In the high-stakes game of evolution and survival, genomes don’t take time to reinvent the wheel. They network, copy, vary, and explore the potential of the information they already hold inside them. As is described in several chapters, genetic change is not something that strikes all parts of a genome evenly. It also has become clear that, as illustrated in Chapter 8 by Evelyn Witkin’s work with sunburned bacteria, the likelihood and type of genetic change, or mutation, can vary depending upon which molecules a cell contains. In other words, rather than being purely passive, the genetic change that a cell experiences can become somewhat conditional on, for example, which proteins the cell itself makes.
CHANCE FAVORS THE PREPARED GENOME
7
Under some conditions, one bacterium may become more open to new ideas, more likely to swallow DNA; a neighbor might feed it a gene that encodes a recipe for destroying an antibiotic. While a genome evolves a balance between faithful copying of itself and exploration through mutation, this is a difficult balance to get right. Perhaps it can never really be “right” because the right balance between fidelity and exploration may change as threats and opportunities in the environment change. As Nobel laureate Barbara McClintock, whose work will be described in Chapter 15, said, “In the future attention . . . will be centered on the genome . . . sensing the unusual and unexpected events, and responding to them.” It is becoming clearer and clearer that some classes of nonrandom mutations are very appropriate to the needs of the organism. In this book, I will discuss recent information that supports this new understanding of evolution, which I first proposed in the technical literature in the early 1980s.4 These ideas attract controversy, but the evidence coming from sources as diverse as the bacteria that cause Lyme disease and our own immune system is growing strong. The work of Richard Moxon and others demonstrates that mutations can become more likely at the very spots in pathogen genomes that speed their race to get a grip on us and to survive. We too are the survivors of many battles with pathogen across the ages, and so those parts of our genome that encode our immune response are creative sites of focused mutation. To reject purely random mutation as the current substrate of genome evolution is not to reject Darwin and Wallace. Indeed, while Darwin’s name may be connected to the phrase “survival of the fittest,” this book emphasizes these words of his: “I have called this principle, by which each slight variation, if useful, is preserved, by the term Natural Selection.”5 Among the variations that I propose are preserved, when useful, are intrinsic variations in the probability of mutation along the genome, as described, for example, in Chapter 4. The flow of information from the biological effects of genetic change to the intrinsic variations in the type, location, and probability of mutation and the mechanisms that generate mutations is not a simple loop, for we are not adjusting an aileron to restore level flight but must survive the unexpected. Still, intrinsic differences in genetic variation would tend to focus in classes of places along a genome where these mutations are more likely to be creative, and tend to move away from areas where changes have done more harm than good. This is a tendency, an adjustment, and not an absolute; for still, now, many mutations do damage. This adjustment would
8
DARWIN IN THE GENOME
emerge because those genomes that accidentally keep losing important information will have fewer descendants across the generations and thus will be less likely to survive. In contrast, genomes will tend to endure when their most likely mutations create effective responses to their most likely challenges, as is detailed in this book. As the fittest molecular strategies emerge through natural selection, by the survival of the descendants of the genomes that encode them, those who remain in the world tend to be those whose ancestors were lucky enough to keep making more creative mistakes. Any genome that we find today, including our own, has been successful, because it has survived through the molecular equivalent of countless canyon gates—though, of course, it has survived only this far, so far. A genome can’t predict what will happen to the next generation, nor can you or I. But genomes have faced some challenges over and over again, such as in host/pathogen battles, and this has left its mark on the genome. A genome evolves a “worldview” of which types of changes, under what types of circumstances, may yield a new function and are less likely to destroy something essential. While a genome can’t predict the future, a genome that has been so prepared by experience is likely to be favored by chance. For me, evolution becomes more conceivable when it is viewed through such a “strategic” biochemical window. The ability to evolve and adapt is an acquired skill, responsive to the environment and acquired through the experience of genomes across generations. Through selection, genome structure emerges from randomness. We live at an extraordinary moment in human intellectual history. Until now, the information that gives rise to all that we try to study in biology and medicine remained hidden from view, as if our eyes were covered by blindfolds. These blindfolds are becoming transparent, and soon they will fall off. The sequences of entire genomes are opening before us. Varied genomes, based on similar chemistry, have spread across the Earth, taking advantage of opportunities, establishing themselves in new environments. After the asteroid hit, what emerged was not another Tyrannosaurus Rex but instead us. Each of us is, in a way, an experiment, and an example of the life-preserving, creative diversity expressed at our moment in time by the human genome. Indeed, we share with one another, no less than with the majesty of the redwoods and the doves, the fact that each of us is a unique creation of the barely tapped potential immanent in the first genomes on Earth.
1 Diversity or Death “We are caught in an inescapable network of mutuality, tied in a single garment of destiny. Whatever affects one directly affects all indirectly.” —Rev. Martin Luther King, Jr., “Letter from the Birmingham City Jail,” 1963
I
t was a mystery. There were very few of these lucky people, but there were a few. They surely had been exposed to HIV, but they simply did not get sick, not even after ten years. Each day each one faced the terror of knowing that the deadly virus had touched his body. And yet each one awoke, morning after morning, to the tentative joy that there was no sign that the virus had gained a foothold inside him. For each one of these lucky few, doctor after doctor—nurses, nutritionists, researchers, reporters, friends—hoped, hypothesized, and investigated. What was he eating, taking, doing differently? Or was it something in his genes? Yet as each was questioned, poked, and wondered over, it remained a mystery. This mystery remained unsolved for a full decade after the HIV virus was discovered in people with AIDS. While this mystery remained unsolved, Ed Berger, at the National Institutes of Health near Washington, D.C., was studying a different HIV mystery. Berger was trying to figure out how HIV gets into cells. HIV does serious damage when it invades the T cells of our immune systems, but it can’t get into just any cell. To get inside a T cell, HIV needs to find a gate it can crash, a receptor on the T-cell surface. The receptor is a protein called CD4. But if we force a cell that does not normally let HIV in to put the receptor protein CD4 on its surface, HIV still cannot get 9
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
10
DARWIN IN THE GENOME
into that cell. So CD4 is not the whole answer. Berger thought that there must be another molecule, a coreceptor, needed along with CD4 to let HIV in. He was right. Berger discovered the coreceptor, called CXCR4.1 CXCR4 allows HIV to spread from T cell to T cell. Finding CXCR4 led investigators to another coreceptor, a protein called CCR5 that looks a lot like CXCR4. CCR5 also can let HIV into cells, but it lets HIV into different cells. Unlike CXCR4, CCR5 is not found at the entry to most T cells. HIV needs to get into T cells in order to destroy the immune system, but first it needs to get into our bodies. CCR5 opens the front door to HIV. For most people, CCR5 is what helps the virus get into the very first cells it infects, where it first touches a person. These cells usually are not T cells. Most HIV particles that attach to CCR5 and infect us cannot get into T cells and thus cannot directly harm our immune system. But, once CCR5 has let HIV inside some of our cells, the deadly invader has penetrated the barrier between being outside of us and being inside us. It has gained a foothold in our bodies; it is living with us, within us. Once inside, HIV experiments. It floats around, poking and prodding, and exploring. It mutates, and these mutations inevitably produce a small change on its surface that it can use to attach to the coreceptor door. Once HIV’s coreceptor binding site mutates to a form that binds to CXCR4, it has evolved the key to let itself into T cells. It continues to mutate and can become an even more unwelcome guest, discovering the keys to additional doors.2 If the HIV in other cells didn’t mutate into a form that could get into the T cells, we might all be infected with HIV without even knowing that HIV exists. This is not a reassuring thought. In fact, I find this haunting: How many other viruses already have found their way inside me, changing, exploring, trying out new keys, perhaps doing no harm—for now? Berger had set out to solve the coreceptor mystery, but his work also explained the mystery of the people who remained healthy after they were exposed to HIV. The discovery of the coreceptors CXCR4 and CCR5 not only explained how HIV gets into cells, but also, unexpectedly, revealed why those few lucky people were so resistant to HIV. The answer to the mystery was that the lucky survivors’ CCR5 coreceptor was different. It was damaged—it was missing a piece, and so couldn’t let HIV in. HIV could not get a foothold in their bodies; it could not get in the door. Out of all the countless molecules in a human body, this one mutation—one small change in one protein—was enough to keep HIV out of the cells of a few lucky people and so saved their lives.
DIVERSIT Y OR DEATH
11
Like all breakthroughs, the discovery of mutant CCR5 answered one question but also led to so many others. How did these lucky people get the unusual CCR5 protein? Obviously it was in their genes, which came from their parents, but what generated the mutant protein in their parents’ genes, or their parents’ ancestors’ genes? Which ancestors were they? Other primates, indeed other mammals, have CCR5 receptors too. Had the ancestors of the HIV-resistant people already encountered HIV, or perhaps another pathogen that uses the same coreceptor to get in? Perhaps smallpox?3 If so, how did their ancestors prepare for it and survive? Or, had the genes of their ancestors, and our ancestors, somehow “learned,” by surviving infection after infection for generation after generation, that something like HIV, another new pathogen, inevitably would come? To become a survivor, it may simply be enough for an individual to be a little different, in a way that provides no apparent advantage in fitness until a new, never before encountered pathogen appears. For an individual, such a difference is great luck; but at the level of the genome, it may reflect a strategy that has emerged in successful genomes, a way to be prepared for the unexpected. The mechanism of evolution, natural selection, usually is described as “survival of the fittest.” But as we look at genomes that have been handed down from generation to generation, what do we mean by survival? What do we mean by fittest? Survival demonstrates fitness: We gather that the successful ancestors, the ones whose genes made it to us, must have had the better genes. But what is “better”? “Better” is a moving target. Continents move; climates change; predators, competitors, food, and the atmosphere all evolve. We call mutations “errors,” but from the perspective of evolution, the most serious error for a genome is to make no mutations. Genomes can prepare for the unexpected by being diverse. If every one of us were the same, and a pathogen were to hit us suddenly and hard (especially before we had laboratories that could work more quickly than generations), we could be wiped out, all of us. The human genome would become extinct in the brief time it took for the pathogen to spread through the human community and do its work. Our libraries and our architecture would be left behind for the product of a future genome to decipher. The reptiles and now the mammals having had their day, perhaps the cephalopods would come next. But, if every one of us were a little different, even in some small way, a few of us might be different in the very way that would protect against this new pathogen. Then a few of us would survive, and with the survivors the
12
DARWIN IN THE GENOME
human genome also would survive. It would not become extinct, but would continue to come forward in time, as it came forward to live within us. It would survive in us, walking, through our descendants, into the future. I think back before protease inhibitors, before antibiotics, before research labs, before we understood how a pathogen spreads (swamp air? drafts?), before we could fight back against pathogens. In that time before worldwide travel, a new pathogen like HIV could sweep through a local human community and kill nearly everyone in it. Indeed, it does not take a journey back in time to see this, just a journey to many places on the planet. HIV cuts huge holes in families, in communities, even in entire countries where there is no access to information about prevention or the right medicine for treatment. As I write this, I am reading a report from Allan Rosenfield, dean of Columbia University’s School of Public Health,4 pointing out that 28 million people in Africa are HIV-positive. There are 12 million HIV orphans; one-third of all adults in sub-Saharan Africa are infected. Parts of Asia are risking the same fate. The human genome will notice this. In an unprotected community, a new pathogen with the right keys could kill everyone except those few who were just a little different, who had a mutant protein in the right place—like, for example, those with the mutant CCR5 coreceptor. After the pathogen had swept through the unprotected community, if the human genome remained there at all, it would have been touched by this tragedy. It would “remember”; it would have been adjusted. Among the survivors, and their children and their children’s grandchildren, the mutant protein would no longer be rare; it would have become the common form of the protein, the one that everyone had. Examining the genome centuries later, we never would guess that it was ever otherwise. We might never be inspired to wonder what protection a now-lost form, or a minor, unnoticed new mutant form, might provide against a different pathogen another day. The pathogen, by removing those without the once-rare mutant protein, would have left behind a changed human genome to be shared by those who survived to live in the future. This has happened before. There is clear evidence that a pathogen has marked our collective genome. This is a pathogen with which we are still doing battle, which kills the equivalent of one to three 747 loads of people every few hours. Most of the passengers on these imaginary 747s are little children, the majority of them under 5 years old. They’ve been bitten by mosquitoes and are suffering from the anemia of malaria, with two-thirds of their oxygen-carrying red blood cells destroyed. Every day, every week, last week, yesterday, today, it continues.
DIVERSIT Y OR DEATH
13
And for each person who dies, many more are ill; 300 to 500 million people suffer from malaria each year.5 It is as if every single person in the United States, plus every single person in England, Germany, and Japan, were infected, each one shaking with the chills and sweating with the fever caused by this persistent pathogen. Of course, few of the real victims are lying in beds in the United States, England, Germany, and Japan. Most of malaria’s direct victims lie in a belt across the middle of the Earth, a belt containing close to half of the people in the world. Their parents do not anchor our evening news, nor edit our daily newspapers, nor write the advertisements that shout to us on TV; but these children share with us this moment in time, in the continued evolution and sculpting of the human genome—and the human genome takes notice. With hundreds of millions of victims across the centuries, including today, the human genome has felt the pressure of the malaria parasite. Of a group of children, friends and brothers and sisters, all bitten by infected mosquitoes, one child comes down with a sudden fever, with seizures, and slips into a coma. Left untreated, half of these ill children die; but even without treatment, half of them live. And some children, even though they are attacked by malaria, don’t get very sick. In the genomes of these survivors, some genes are different. Just as HIV kills those who lack the mutant CCR5, leaving as survivors those with the rare version, malaria too has sculpted the human genome. This sculpting of our genome, this battle with malaria, is not without its victims. For these victims, too, the struggle is painful. Tears roll down the cheeks of an infant screaming in pain. She cannot explain it, but it feels as if countless nails are being driven through her body. Her red blood cells, the carriers of desperately needed oxygen, are bent out of shape and are jamming up her blood vessels. Her nerves, sensing the lack of oxygen, are exploding in pain. There were over 300,000 of these children, born last year,6 in a tropical band across Africa and in families whose ancestors lived in these places, suffering from what we have come to call sickle cell disease. What happened to leave so many children suffering in, of all places, the paradise of our imaginations, the tropics? Biochemically, at first glance, it is a minor thing that is causing so much suffering for the little girl—a change in a single amino acid in one protein out of many tens of thousands of distinct proteins in her red blood cells. But this tiny change is in hemoglobin, a vital molecule that carries oxygen to all of our tissues. Surely such a mutation, one that messes up a vital molecule, damages the red blood cells, and causes so much pain, should have been removed from
14
DARWIN IN THE GENOME
the human genome by natural selection long ago. But it has not been removed. In fact, this painful mutation seems to have thrived and spread. This child’s pain comes from the wounds of a battle the genome has fought against a tiny adversary, invisible to her unless someone were to give her a microscope. The tiny adversary is a protozoan, with its relatives called Plasmodium, that causes malaria. Both copies of her hemoglobin gene are altered, and the mutant protein encoded by these genes has sickled her red blood cells out of shape and jammed up her tiny blood vessels; but when only one of the copies of this hemoglobin gene is altered, as in the genome of each of the little girl’s parents, it can be a good thing for people who live at sea level in the tropics, for it protects them against malaria. And so, as malaria continues to take down its jumbo-jet loads of people in the tropics every few hours, those who survive the attack are more likely to have a mutant hemoglobin within their red blood cells. In the tropics, the malaria parasite has left its mark on the human genome. For malaria and mammals, the oxygen-carrying red blood cells have been a major theater of battle. The hemoglobin molecule, the molecule that catches oxygen in our lungs and delivers it through the body, has been a major battleground. Not only sickle cell hemoglobin, but also the thalassemias, a swath of variants in hemoglobin structure and regulation, follow the anopheles mosquito that carries malaria. Surely, after all this time, the mammalian genome should have won; it should have evolved a way to keep the malaria parasite out of our cells. Indeed, we have changed. But so has our adversary. As the malaria parasite has sculpted our human genome, we surely have sculpted its genome, too. We can change, but there is a limit to how drastically we can change our red blood cells. We can fool around only so much with our life-sustaining oxygen-carrying mechanism. And so the battle with our ancient adversary continues. Perhaps now that we finally are gaining the real-life equivalent of the magical secret decoder ring with the sequencing of the complete malaria genome, we will find a more effective vaccine or therapy and at last overcome malaria’s bag of tricks. Perhaps our new level of consciousness will finally bring us complete victory over our ancient adversary. But not yet. For now, malaria continues to take down its 747 load of victims every few hours, last week, yesterday, and today, this afternoon. If you happen to glance at your watch again in a few hours, note that another planeload of children will be gone. As we look into the forest at night, instinctively we fear snarling fangs and coiled vipers. But it is the once-invisible predators—protozoa such as
DIVERSIT Y OR DEATH
15
malaria; bacteria such as the plague, typhus, and tuberculosis; and viruses such as smallpox—that take the largest toll on our species. Just as they still threaten us today, these tiny predators put selective pressure on our ancient ancestors. As our ancestors escaped from and banded together to capture the large predators, the predators we listen for when we wake up in the moonless forest at night, our genes have been doing battle with the tiny ones. And new ones keep emerging. Where does Ebola hide deep in the forest? One evening in the mid-nineteenth century, Alfred Russel Wallace lay on his cot at the edge of the rain forest shivering with fever. He had traveled halfway around the world to study nature, and now, on an island in the Moluccas, he was getting a much closer view of it than he had planned. As he lay on his cot for hours each day, alternately shivering and sweating, he later wrote, “I had nothing to do but to think over any subjects then particularly interesting me.” One day he began to think about Malthus’s book Principles of Population, which he “had read about twelve years before.” Wallace began to think about Malthus’s “clear exposition of ‘the positive checks to [population] increase’—disease, accidents, war, and famine.” He connected the idea that disease, accidents, war, and food shortages, which would limit the ability of the human population to keep increasing, also could limit the increase of animal populations.7 Animals breed so quickly, he reasoned, that there must be an enormous loss each year or “the world would long ago have been densely crowded with those that breed most quickly.” As he reflected on “the enormous and constant destruction which this implied,” his mind focused on the question, “Why do some die and some live?” This was not a purely abstract question at that moment for Wallace, who was lying with a raging fever in a hut on an island far from home. For him, “the answer was clearly, that on the whole the best fitted live. From the effects of disease the most healthy escaped; from enemies, the strongest, the swiftest, or the most cunning; from famine, the best hunters or those with the best digestion; and so on. . . . That is, the fittest would survive.” Wallace’s next thoughts were about dramatic changes in the environment and the great amount of individual variation there is within a species. Then at once I seemed to see the whole effect of this, that when changes of land and sea, or of climate, or of food-supply, or of enemies occurred—and we know that such changes have always been taking place—and considering the amount of individual variation that my experience as a collector had shown me to exist, then it followed that all the changes necessary for the adaptation of the species to the chang-
16
DARWIN IN THE GENOME
ing conditions would be brought about; and as great changes in the environment are always slow, there would be ample time for the change to be effected by the survival of the best fitted in every generation.8 With his own life threatened by disease, Wallace had discovered the idea of natural selection, survival of the fittest, all on his own, before hearing the idea from Darwin. While the phrase “survival of the fittest” seems to ring in our ears, there is another, quieter phrase that needs some attention. Wallace said that his realization of the role of natural selection was based on “considering the amount of individual variation that my experience as a collector had shown me to exist.” Variation is an essential prerequisite to selection. Variation comes from mutation: changes in a single genomic letter, changes in a block of letters, and the rearranging of pieces of DNA. The ultimate creative step in variation, for those who have survived to become adults, is the cutting and pasting and mixing and sorting with the genome of another adult that happens when we create a new life together—a child created by mixing pieces of the genomes of all our ancestors. Survival-of-the-fittest: The words are spoken as one word. But what carries our own specific DNA forward in time? Our DNA survives in a partnership with others, carried like candlelight passed through a branching chain of individuals. As the torch is passed at the border of each new generation, our DNA gets cut up and mixed with DNA from other people. The work of the other genes in the new mixture may increase or decrease each gene’s “fitness.” Ten generations from now, an eyeblink in evolution, if your DNA has survived the journey, it will have been diluted to a mere thousandth of your descendants’ genomes. It will be mixed in with the genomes of thousands of other people who are now alive, most of whom you do not know. Perhaps a part of your DNA will find itself connected to the DNA of someone in the blue car, in the center lane that you passed yesterday morning, or that person who stepped aside as you left the train in the subway, or the child you saw in the arms of his mother, clinging to a tree, in a flood in a distant land reported on the evening news. Together, at a time in a future that we cannot clearly imagine and that we will not ourselves live to see, all of you may share in the creation of a child. With luck and fitness, these descendants will survive. Is this how, and when, we tell that you were the fittest? The fittest what? The fittest when? Now, or when future fragments of your genome encounter a new pathogen? What is “better”? What is fit when you are aiming at a moving target? Fitness emerges as a strategy, not a goal; a process, not a place.
DIVERSIT Y OR DEATH
17
The pain and tears of the little girl with sickle cell anemia are her wounds in a struggle that is for us, for the survival of our shared genome. She is suffering because her parents carry important information in their genes, information to be shared, through her parents’ children’s children, with our children’s children, information that protects against malaria. For even today, malaria is not tamed, and it is not caged: In 1999, at a camp about a 11⁄2 hour drive from the Empire State Building, two 11-year old boys got malaria. To survive, we must absorb this truth, that our genome’s ability to change, to explore, to incorporate the discoveries of many individuals—and the diversity that results from this exploration—is a central part of our fitness. It is a lesson incorporated into our genome through billions of years of evolution. If we do not treasure human diversity, we risk an eternally broken chain, the end of our species, the loss of our future. If the human brain and the human heart do not learn fast enough what we have known in our bones, and in our eggs and sperm, for millennia, the genes that have brought us this far, through all of their diversity, will be wiped out. Our hopes, our future, indeed our very survival in a distributed, gene-mixing rebirth, depend upon our connections with one another. We are all, profoundly, siblings in the present, parents of the future.
This page intentionally left blank.
2 The Magic Staircase In this way an infinite variety of nucleotide [letter] sequences would be possible, to explain the biological specificity of DNA. —Rosalind Franklin1
E
ach time a skin cell divides to make two new cells or parents pass their DNA to a child, the information stored in the DNA must be copied very carefully. When I stop to think that the information in our DNA is stored in 3 billion pairs of pieces linked together in 46 strings, one string to a chromosome, I do not wonder that mistakes can be made in copying this information. In fact, I wonder why there aren’t many more mutations. To copy so much, so carefully, takes precise machinery, as well as backup machinery that can fix essentially any error. You might envision DNA as a carefully preserved reference library. The instructions needed to make a hand, an eye, an ear, a virus, a tulip, a bee, or a person are stored there, in code; so are the instructions for making molecular machines, including molecular machines that copy DNA and machines that can proofread their copy. Included in the DNA reference library, too, are the instructions for making a molecular machine that can decode the information encoded in a molecule of DNA and use it to build molecules of proteins. When DNA is decoded, its information first is copied into another molecule, RNA, that travels like a messenger to a protein factory, where the information the RNA carries is translated into the strings of amino acids
19
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
20
DARWIN IN THE GENOME
that we call proteins. Each amino acid in the protein is specified by a string of three letters in the DNA and RNA. For example, the letters CAG encode the instructions to put the amino acid Q into a protein.2 In an English-language library, information is stored using an alphabet of 26 letters, A through Z. In a DNA-language library, information is encoded using an alphabet of four letters, A, G, T, and C. Of course, these letters are not really letters, but chemicals called adenine, guanine, thymine, and cytidine. A and G are similar to each other, and are bigger than T and C. These four letters are strung together as if in words—GACGCTAGATGGATTAAC. . . . For each protein in our bodies, there is a place in the DNA, which we refer to as a gene, where the order of these chemicals or letters specifies the order in which to connect amino acids in a string that will fold up to become that protein. A protein molecule may be part of a larger structure, like the fibers in our muscles, or it may be an enzyme, a catalyst that makes things happen much more rapidly than they would happen in a lifeless pond. We may envision these genes being brought out by their cellular keepers, as if on velvet from a locked vault, for precise reading of their instructions—for faithful copying—and handed down with great care from parents to children. We may envision this precious DNA being carefully protected from dust, and certainly—to the greatest possible extent—from mutation. But maybe the peaceful library is not such a good way to describe DNA, because DNA is not like the motionless double helix of textbook covers. Living DNA is active, dynamic, and busy. You can’t assume that if you’ve seen one patch of double helix, you’ve seen them all. If it were a creation of performance art, a tourist attraction on a grand plaza, all guidebooks would recommend DNA as a “must visit.” We would crowd around, look up, and be captivated. We would see two strong ribbons, twisted about each other to form a double helix, extending far up from our imaginary plaza. Running up the center of the double helix, as if the two helical ribbons were banisters, is a spiral staircase (see Figure 2-1). The steps of this staircase are all the same width and come in two color schemes: one part gold and part cobalt blue, the other part turquoise and part apricot. At a scheduled time, starting from the top, the gold part of each step separates from its cobalt partner, and each turquoise patch separates from its apricot partner. Starting from the top, the two ribbons twirl and move apart. In the midst of the twirling, separating ribbons, two new helical ribbons begin to grow, one beside each of the original ribbons. To keep things from getting tangled, a ribbon is cut, its partner helix passes through it, and the ribbon ends are reconnected, precisely and quickly. Over and over, they
THE MAGIC STAIRCASE
21
Figure 2-1 The magic staircase. Two strands spiral around each other, forming a double helix. Each strand is a string of the letters A, T, G, and C, connected by linkers. The information that DNA carries from generation to generation is encoded in the order of the letters in each strand. When DNA is duplicated, the strands separate; each strand specifies the order of letters in its new partner by the rule that where there is an A on one strand, there will be a T on the other; and where there is a G on one strand, there will be a C on the other.
are cut, twirl, and reattach, in an intricate choreography. Attentive spectators begin to notice a recurrent pattern. It becomes obvious that you can predict the next color that will join a growing ribbon by looking at the strand of the original ribbon beside it. If there is a gold patch on the original ribbon, you can predict that it will attract a cobalt patch to form part of a new gold/cobalt step connecting the two ribbons; cobalt will attract gold, apricot patches will attract turquoise, and turquoise will attract apricot. Each of the two original strands is used
22
DARWIN IN THE GENOME
as a template to direct the order of the colors in its new partner strand. When the swirling stops and two new double helices are before us, the color scheme in each is the same as that in the original double helix; the two new helical staircases are copies of the original, with the same twocolor steps, gold/cobalt and apricot/turquoise. We may watch, enthralled, as the beautiful sculpture spins in its complex choreography. But we know that there is more to the sculpture than its form and the patterns of its dance. The most captivating part is that messages are encoded in the sequence of cobalt, gold, apricot, and turquoise— C, G, A, and T. If we can break the codes, we can decipher extraordinary messages. These ribbons encode a life and create a foundation on which future generations can emerge. In life, when a DNA double helix unravels to be copied, and the two new double helices grow alongside their partners, a rapidly acting copying machine, composed of proteins, has the job of making a careful match at every step. Some of the proteins in the machine are enzymes, the catalysts that attach the letters to each other. Some of the proteins hold the machinery together; others hold the machine to the DNA as it moves along. The copying machine has to be careful, but it also has to be fast. To copy the 3 billion steps in each of our cells each time the cell divides takes about 8 hours, with each of about 1000 machines attaching about 80 new letters every second. One tiny missed match among the billions would be a mutation; the child would be different from the parent. As the copying machine builds the chain, there are two steps to think about. First, the right letter must float into the machinery to become the new partner of a letter on the template strand (A floats in to pair with T, T to pair with A, G to pair with C, and C to pair with G). Once this new letter is in place, the enzyme creates a strong link to make the new letter part of the growing DNA chain. Like beads in a necklace, the letters in DNA hang on linkers that attach them to the chain. DNA’s linker is in part a sugar, deoxyribose, the D in DNA. To build the growing chain, the enzyme attaches the deoxyribose linker of the incoming T or A or G or C to the deoxyribose linker of the T or A or G or C that the enzyme has just attached to the growing chain one step before. How does the machinery get the correct letter into each step of the template strand? If a G floats into the empty space across from a T instead of T’s correct partner, A, the shape of the step will be wrong. Because the G doesn’t fit right, it tends to float away, leaving room for an A to float in. In contrast, if an A floats in, it fits quite comfortably, and thus
THE MAGIC STAIRCASE
23
it is more likely to stay long enough to be attached to the growing chain. In fact, when this was measured,3 when copying a T, A was about 1000 times less likely to float away than G was. In DNA, the letter G (or guanine) is the closest in shape to A (adenine). When a C or another T floated in beside the T that was about to be copied, it was even more likely to slip away. Because the correct letter fits the best and stays in the machinery the longest, it is most likely to become part of the growing chain. Not only does the wrong letter tend to slip away, but also, if the letter pair is the wrong shape, the copying enzyme seems to have a hard time reaching around the back to attach the new letter to the growing helix. In fact, it took four times as long to attach a G across from this T as it did to attach an A, and it took eight times as long to attach another T or a C. Because it takes longer to attach the wrong letter, and during all that time the wrong letter has a high chance of slipping away, mistakes are infrequent. But whether it is copying a tiny genome or a human genome, occasionally the machine does make an error. In fact, in the example described above, a T-G pair was formed instead of a T-A pair once every 20,000 times the T was copied. T-C and T-T pairs were each formed only once every 500,000 times the T was copied. If I made only one mistake every 500,000 times I made a decision, I would be pretty proud of myself. For DNA, however, this is good, but not good enough. In our two-stranded 3-billion-letter genome, one mistake every 20,000 to 500,000 times would be 12,000 to 300,000 mistakes every time one of our 50 trillion cells divides. At that rate, our genome would not come close to making it to an adult from a fertilized egg. In fact, our DNA copying machinery does much better than one mistake every 20,000 to 500,000 times, because it actually can proofread its work and, when it does insert the wrong letter, correct it on the fly. It can break the link it just made and get rid of the mistaken letter, the wrong partner, catching errors so fast that its accuracy is improved another 1000 times—it removes 999 of every 1000 errors. Still, some mistakes remain after the copying machinery has moved on. One team looked at 414 mutations that were left after copying many examples of a 210-letter region of a gene in bacteria grown in their laboratory. There were different kinds of mistakes. In 7 of every 10 mistakes, the wrong letter was attached. Nearly half of these wrong-letter mistakes were of the same kind—a T, rather than a C, was paired with a G. Nearly 1 in 5 mistakes were deletions, where a whole patch of the DNA was left out. (More than half of these deletions occurred at one spot, over and over.) Nearly 1 in 10 mistakes involved the insertion of a patch of
24
DARWIN IN THE GENOME
extra letters, and about 1 in 20 of the mistakes involved adding or removing a single letter.4 These copying errors are not left in the DNA, for there is yet another level of protection for the messages in DNA. A “mismatch repair” machine follows to correct the remaining mismatches between the template letters and their new partners. In the 210 letters of the bacteria’s gene discussed in the last paragraph, mismatch repair corrected 996 of every 1000 errors, leaving only 4 of these errors behind. Just as some kinds of mutations happened more often than others, mismatch repair fixed some kinds of errors more reliably than it fixed others. It kept Cs from replacing Ts seven times better than it kept Ts from replacing Cs.5 The mismatch repair machinery is especially careful to repair places where extra letters have been inserted or letters have been deleted.6 Insertions and deletions generally are much more serious than a change in a single letter, because a deletion or insertion throws the whole message out of frame: mngitu ni ntelligibletot he rea der. But still, after all of this attention to detail, and the care taken before making the bond, and the proofreading and repair of mistakes, out of the 1000 errors that were left after careful copying and proofreading, 4 still remained. Sometimes the DNA at a spot is badly damaged even before it is copied, perhaps by radiation or something else in the environment. It might be so badly damaged that the careful copying enzymes can’t even recognize it. Was it once a T? When the machinery runs into a damaged letter, it simply stalls there, leaving a cell with a partly copied piece of DNA. What to do? Call in another DNA copying enzyme that can make a quick and dirty patch, so that the DNA copying machinery can move on.7 This copier is not so demanding about the proper shape, so in an emergency it is able to attach something opposite the damaged piece and allow the copying to pass the block. This less finicky copier moves slowly and can be replaced by a more careful enzyme once it gets past the mistake. Later, other repair systems may be able to fix up the sloppy repair job. If not, where the DNA was damaged, for example by ultraviolet light, a mutation is left behind. Just to copy the genome of the relatively simple bacterium Escherichia coli requires complex machinery. Its copying machine is made of 10 different kinds of proteins, some in more than one copy, so that there are 18 pieces to the machine that we know of so far. In addition to the protein enzymes that actually hook the new letters to the growing strand, there are the proofreaders, the proteins that hold the machinery together, and
THE MAGIC STAIRCASE
25
the proteins that hold the machinery on the DNA so that it doesn’t float away. What a job. And this happens over and over again, each time the DNA is copied; this bacterial machinery copies 500 letters per second. When the copy is faithful to the original, the message will be preserved; when an error is made and not caught in time, the message will change; there, the new cell will differ from the original. With the combined care of the copying, proofreading, and repair machinery, there is on the average only about one unrepaired error in every billion to 10 billion letters copied—about one mutation every time one of our cells divides. This mutation may have no effect, it may be bad, or it may even be a good thing. If it happens in our skin cells, we may never notice it, or it could lead to skin cancer. If it happens in what become sperm or egg cells, when we are passing DNA to our children, it is the substrate of evolution.
This page intentionally left blank.
3 Predators Battle Prey in the Genome And we too are impressions left by something that used to be here. —Stephen Batchelor1
A
shell caught the athlete’s attention as he walked along a Pacific beach. It had an intricate, beautiful pattern, and it was large for a shell. He picked it up, and, looking closely at the treasure in his hand, scratched it with a pocketknife. Unfortunately, this shell was occupied, and the young man’s scratching at the opening was interpreted as predation. The occupant reacted to defend itself. “Hey, I think this thing stung me,” the man said. According to his mother, he felt no pain, just some numbness, but soon could not move his legs; he slipped into unconsciousness and, within 5 hours, he was dead.2 I learned about this deadly cone snail from Baldomero Olivera, a biochemist, who made good use of some free time he had when he returned to his native Philippines. He was waiting for equipment that he needed for his research on DNA to reach him from overseas. Olivera decided to use that waiting time to satisfy his curiosity about how cone snails, which come in many sizes and beautiful, collectable, patterns, kill their diverse prey.3 By the time his equipment arrived, he was hooked on the snails. Life on a reef is both rough and biochemically competitive. The beautiful killer cone shell contained a snail that lives by fishing. A cone snail can catch a fish. True, the snail would lose any race against a fish through the 27
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
28
DARWIN IN THE GENOME
water. But after all, we do not catch most fish by swimming after them either. Some cone snails use a strategy that starts a bit like ours: The snail floats a “worm” in the water. This is not, of course, an actual worm; it is a part of the snail’s body that looks like a worm. When the fish bites the “worm,” the snail injects a cocktail of nerve and muscle toxins into the fish that, essentially, short-circuits it. The fish can’t move, and the snail has its meal. Olivera, working together with Lourdes Cruz, a professor in the Philippines, found more than one toxin in cone snails—many, many more. Each of the 500 cone snail species has at least 50 and sometimes even as many as 200 different toxins. Different species of cone snail kill different types of prey—fish, worms, even other kinds of snails. They have a cornucopia of toxins with a wide range of toxic effects, interfering with different types of nerve signals. Cone-snail evolution was beginning to look like a toxin-generating machine. Olivera began to toy with the idea that cone snails might have evolved a genetic strategy for generating diverse toxins rapidly. If cone snails were able to evolve such a genetic strategy, it would give them an advantage in molecular evolutionary wars. A large and flexible repertoire of new toxins would enable the cone snails’ progeny to adapt rapidly, across only a few generations, to changes in prey, predators, and competitors. In a fast-changing competitive environment, it is good to have a wide range of options. Olivera’s thinking about the evolution of cone-snail toxins began to focus on a mismatch between something everyone expected and something he observed. When two things that always match suddenly don’t match, pay attention. There is something unusual afoot. The standard explanation as to why cone snails have so many different toxins would be that the snails’ DNA varied randomly and the snails that survived were the ones that got it right. This standard explanation began to unravel when Olivera noted that most of the variation between different toxin genes was focused in one corner of the gene. At first, it looked as if the standard “random mutation, then selection” theory could explain this focused genetic variation. Random mutation generates changes throughout the gene, this explanation says, but changes in most parts of the gene provide no selective advantage. If changes in a particular part of the gene provide no advantage, then a snail with those changes is no more fit than the starting snail, and so there is no reason to expect an increase in snails with changes in that part of the gene. Changes in some parts of the gene actually hurt the new snail, so those snails have a disadvantage; we won’t see many of them or their children either. Changes in the side of the toxin gene
PREDATORS BATTLE PREY IN THE GENOME
29
that creates new toxins may actually help the snail survive, and so we get a lot of snails with changes in that corner of the toxin gene. So, the standard theory explains, we see more changes in one part of the toxin genes because these changes generate new, improved toxins; changes that generate new, improved toxins would be the only changes favored by selection; changes anywhere else in the gene might not have been selected for—or might even have been selected against. Thus, random mutation and selection would explain why one region of the toxin gene has more changes. The standard explanation says that the toxin DNA only appears to change more quickly; change actually occurs randomly everywhere in the genome, but we do not see the other random changes because the snails that inherit them do not have any advantage. The problem with the usual explanation was that only some of the frequent changes that Olivera found in one region of the toxin gene occurred in places that actually change the toxin protein itself. If many of the frequent changes in the toxin DNA were not changing the toxin protein, these specific changes couldn’t be selected for by selecting for a better toxin. So why were these changes so much more common than changes in other parts of the DNA? Something other than selecting for better toxins one at a time must have been attracting variation to the high-mutation corner of the toxin DNA. If genetic variation is more likely to happen in one corner of the DNA than in another, mutation isn’t random. Before saying any more about this, I need to explain how the DNA that codes for the toxin can change without changing the toxin protein. Mutations in protein-coding DNA can occur without changing the protein itself because the genetic code is degenerate. In code language, degenerate means that there is more than one way of saying the same thing. Because more than one sequence of DNA can code for the same toxin protein, it is possible to change the sequence of letters in the DNA without changing the toxin protein at all. A protein is born as a necklace, a string of amino acids. Amino acids are added to the string in the order specified in the genome. This order is specified by reading DNA in blocks of three letters, or codons, which is why the genetic code is called a triplet code. Each of the three letters in a codon can be any one of DNA’s four letters, and thus there are 4 × 4 × 4 = 64 possible codons. The 20 letters shown across the top of the boxes in Figure 3-1 represent the 20 amino acids that form our proteins. They are shown once each in alphabetical order only for illustration. The number of times each amino
30
DARWIN IN THE GENOME
A GCU GCC GCA GCG
C UGU UGC
D GAU GAC
E GAA GAG
F UUU UUC
G GGU GGC GGA GGG
H CAU CAC
I AUU AUC AUA
K AAA AAG
L UUA UUG CUU CUC CUA CUG
N AUU AAC
P CCU CCC CCA CCG
Q CAA CAG
R CGU CGC CGA CGG AGA AGG
S UCU UCC UCA UCG AGU AGC
T ACU ACC ACA ACG
V GUU GUC GUA GUG
W UGG
Y UAU UAC
stop UAA UAG UGA
M AUG
Figure 3-1 The genetic code.
acid is used, and the order in which they are connected to each other, is what gives different proteins their different properties. The triplets listed under each letter are the codons that can be used to encode each amino acid. For example, the codon GAG means “Add the amino acid E to the growing protein,” and GAC instructs the proteinassembling machinery to add a D. Because messenger RNA has an alphabet of four letters, G, C, A, and U (RNA uses U where DNA uses T), there is a total of 64 codons (including the “stop” codons used to mark the end of the protein). With only a few variations in meaning, these same codons are used by all life on Earth. By choosing different combinations of the codons available to encode each letter, there are 4 × 2 × 2 × 2 × 2 × 4 × 2 × 3 × 2 × 6 × 1 × 2 × 4 × 2 × 6 × 6 × 4 × 4 × 1 × 2 or nearly 340 million ways to encode this particular string of 20 amino acids (times three ways to encode “stop”). There are 64 codons, but, typically, proteins are strung together from an ingredients list of no more than 20 amino acids. Thus, there are more than three times as many codons as there are amino acids; 64 codons specify 20 amino acids. More than one codon is available to encode each amino acid, allowing some flexibility in the use of language and enabling the code to be degenerate. Indeed, most (but not all) amino acids are encoded by more than one codon; some amino acids have as many as six codons. In other words, for some amino acids, the DNA can use any one of six codons to signify that this amino acid should be added to the growing protein. And if different codons can code for the same amino acid, then different strings of codons can code for the same protein: Either AGCGAAGAGGAT or
PREDATORS BATTLE PREY IN THE GENOME
31
TCTGAGGAGGAC could encode the protein piece “SEED.” Your DNA can say it either way. If the cone-snail toxins did change just by random mutation followed by selection for the “winner” toxins, we would expect to see a lot of changes in the DNA that were not synonymous; we would expect to see changes that would change the amino acids, like changing the amino acid E to D by changing the codon for E, such as GAG, to a codon for D, such as GAC. We certainly wouldn’t expect to see a lot of extra changes that only replaced a codon with its synonym in the toxin genes, such as changing a codon for D to another codon that still means D. Of course, there would be some synonymous changes just from random mutation, but the rate of change between these synonyms wouldn’t, randomly, always be higher in the same corner of the toxin gene than they are somewhere else in the cone-snail genome. What we would expect, assuming that mutation is random, is not what Olivera found. When he compared DNA from many different cone-snail toxins, he found a very high rate of change between codons, including a high rate of change between synonyms, in the region of DNA that codes for the toxins. He also found that this part of the gene often gains and loses patches of DNA. Selective pressure might favor changes in codons that change the amino acids to create new toxins, but why should there be so many changes, even between synonyms, in this small region of DNA? Olivera wondered if a special variation-generating mechanism might somehow be directed to the little stretch of DNA that codes for toxin proteins.4 Perhaps selection has favored not only snails that evolved a good new toxin, but, most significantly, snails that evolved a strategy of directing genetic change to the stretch of DNA that codes for the toxin. Rather than mutating and waiting for selection to assess each change letter by letter, some snails may have been favored by evolving mechanisms that focus mutation in this important corner of the toxin genes. Some of these changes would be synonyms, and so would not affect the protein; but many of these rapid changes that were directed to the toxin region would be changes in the toxin protein itself. This mechanism would guarantee that generation after generation, cone snails could retain a competitive edge by being able to try out new toxins rapidly. The snails could, of course, evolve new toxins quickly simply by having a generally high mutation rate, and perhaps some now-lost cone snails used that strategy. But why mess up a lot of good genes when it is, generation after generation, specifically toxin diversity that you need in order to help ensure that some of your progeny will survive changes in predators, prey, and food? The progeny of cone snails with uniform, higher mutation rates would be
32
DARWIN IN THE GENOME
weakened by errors in important housekeeping genes. (On the other hand, the progeny of cone snails with uniform, low mutation rates might not be able to defend themselves against a new competitor or a new predator, or they might miss the opportunity to catch a new species of fish if old food sources became scarce.) It is more likely that among the progeny of cone snails with a high mutation rate that was focused specifically on the toxin gene, there would be some progeny with new toxins that were useful when there was a change in their predators or prey, or in competition against other snails. The progeny of these snails would have an advantage and would survive. It is a more efficient route of evolution, and a clever genetic strategy, to increase genetic exploration for a toxin and let the rest of the genome change more carefully. So perhaps there has been positive selection in the evolution of the cone-snail genome for a strategy that generates new toxins quickly. Other predators, too, from vipers to scorpions, seem to have special mechanisms to generate rapid variation in the very genes they need to attack their prey. Cone snails have many, many progeny. It is likely, but it has not yet been proven, that cone snails with a modulated mutation rate that is especially high in the toxin gene and lower elsewhere would be the winners, and thus would be the ones that we admire while we step gingerly around them on Pacific beaches. We can step around the cone snails on the beaches and vipers on the trails, and shake scorpions out of our shoes in the morning, but there are other dangers that are harder to avoid. They cannot be avoided by staying home; they lurk on our cutting boards, grow in our sponges, attack from the air, travel quietly in unseen insects, hide in our food, and sneak across in a handshake. Many of these tiny predators, pathogens that are individually invisible to us without a microscope, evade our immune system by quickly changing the genes that encode their coats. They too appear to have come up with something better (from their point of view) than random mutation. Just when our immune response gets ready to kill anything wearing the pathogen’s coat, many pathogens can put on a new coat. Bacteria that look like the spiral that twists around a screw are called spirochetes. The spirochete that causes Lyme disease has a large selection of patterns for its coat. We can watch what happens by taking blood samples from mice infected with Lyme disease spirochetes. Each time the spirochete bacteria that cause Lyme disease are isolated from infected mice, a large patch on the spirochete’s coat is different.5 The spirochete keeps changing its coat by inserting new patches of DNA into its coat-protein gene; each new patch of DNA results in a different patch of amino acids in
PREDATORS BATTLE PREY IN THE GENOME
33
the coat protein, as if one day the back of the coat is red, the next day green, and the next day orange. The mouse’s immune system, which was alerted to grab anything wearing a coat with a red patch and then has to learn to grab anything in a coat with a green patch, and so on, can’t keep up with the pathogen’s wardrobe. The Lyme spirochete can thus continue to live within the mouse, surviving in spite of the mouse’s stimulated, searching immune system. The Lyme spirochete takes its many quick wardrobe changes from a large inventory of extra DNA pieces called plasmids. It can also mix and match pieces of the stored coat patches, making patchwork of the patches and giving them even greater variability. In fact, it is estimated that the spirochete has 1030 (1,000,000,000,000,000,000,000,000,000,000) different patch patterns. If we examine the spirochete’s coat carefully, we can see that surrounding the ever-changing patch in the spirochete coat protein, there is a completely conserved repeat of five amino acids, EGAIK. If we stop to examine the 15 letters of DNA that, using the triplet code, encode these five amino acids, we find something very unexpected. Because of the degeneracy of the genetic code, there are nearly 200 DNA sequences that could have been used to encode the five amino acids EGAIK. But the 15 DNA letters that encode EGAIK in the repeated patch in the spirochete coat always are part of an identical sequence of 17 letters, TGAGGGGGC TATTAAGG, that is found on both sides of the repeat and in all samples of spirochete DNA tested. We never see TGAAGGAGCAATAAAAG or TGAGGGCGCCATCAAGG or any of the nearly 200 other ways to encode the very same amino acids. If these really are synonymous codons, why is it that time after time, spirochete after spirochete, it is always, always TGAGGGGGCTATTAAGG? The other 192 synonyms for EGAIK are not acting as synonyms should. You and I could make a plan: If you say “I’ll call you tonight,” it means that you’ll call me at 8 P.M., but if you say “I’ll phone you tonight,” it means that you’ll call me at 9 P.M.. With two different ways of saying the same thing, we can create another code that sends more information underneath the message. People who didn’t know about our code would hear you say, “I’ll call you tonight,” and think they knew the whole message. But you would have transmitted an additional message, with even the existence of the message remaining hidden from those who did not suspect it was there. Might there be a secret message that attracts genetic variation to a patch of DNA? Some kind of message would in fact be needed to attract enzymes that cut out the DNA in between the spirochete’s repeats and
34
DARWIN IN THE GENOME
replace it with DNA that codes for a different patch of spirochete coat and is stored somewhere else in the spirochete’s DNA. When you have an enzyme that cuts and pastes DNA floating around, you don’t want it to cut just anywhere. It is safer to require strict recognition of a special sequence, an address, to direct the cut to the right place. In the Lyme spirochete, the cut could be placed precisely on either side of the ever-changing patch in the coat-protein DNA by a message hidden in the DNA. This targeted cutting allows the Lyme spirochete to make its rapid wardrobe changes without waiting for random changes everywhere in the spirochete’s DNA to make a lucky hit. If the DNA that encodes the amino acids EGAIK changed to any one of nearly 200 synonymous sequences, it still would encode the amino acids EGAIK, but the DNA recognition site would be damaged, the biochemical mechanism that exchanges pieces of DNA wouldn’t work, and the spirochete couldn’t keep changing its coat. Those spirochetes that tended to make what appear to be synonymous changes in the DNA that encodes EGAIK would in fact lose the ability to change their coats and could be caught by our immune response. Perhaps codons that have been considered synonyms ever since the genetic code was cracked in the 1960s do not always behave as true synonyms. True, the codons are synonyms in that they code for the same amino acid. But what we thought were synonymous codons may have additional, distinct, content-dependent meanings for the proteins and enzymes that cut and paste DNA. The spirochete coat protein has a telltale pattern: conserved patches surrounding a region of great variability. This pattern may point us to a broader biochemical strategy that the successful spirochete has evolved for protection from its most predictable challenge: the host immune system. In fact, if you look very closely at the spirochete’s genome, you will find other stretches of absolutely conserved DNA around patches of changing protein. We expected random codon choice, but codon choice is not random in the spirochete coat protein. Something unusual is afoot. There is extra information embedded in the stretch of DNA that encodes the amino acids EGAIK. The spirochete genome will call us at 9 P.M.; it has something else to say.
4 Mutation Is Not Monotonous Rare is not a synonym for random. —Lynn Ripley1
F
or all the reverence in which we hold it, DNA is just a molecule, composed of chemical pieces that we represent by the letters A, T, G, and C. The information encoded in one DNA double helix is different from the information encoded in another DNA double helix with a different sequence of As, Ts, Gs, and Cs. The different order of the pairs of letters on the strings makes each of us different, and it makes the DNA itself different, too. Looking from one side of the double helix, there are four letter pairs: A on this strand paired with T on the other, T on this strand with A on the other, G on this strand with C on the other, and C on this strand with G on the other. Each of these pairs looks slightly different from the others. As DNA is run through the copying machinery, these differences in the DNA make some mistakes more likely to occur than others. The proteins that hold onto and copy DNA interact with it as one molecule with another molecule. Physical forces dictate their interactions. The two strands of the long helices seem to breathe, spreading apart a little, then settling back together. How often each step in the helix breathes depends upon how tightly each pair of letters holds each step together. The strength of an A-T pair is different from the strength of a G-C pair; thus the A-T pair’s tendency to separate slightly for an instant—the breathing of the helix—is different from the G-C pair’s. The structure of DNA
35
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
36
DARWIN IN THE GENOME
changes subtly, then, as we move along the helix. The differences in DNA sequence that carry the linear code also affect the amount of breathing. How much DNA opens up at one step depends not only upon the letters at that step, but also upon the letters nearby. A string of six A-T pairs, one on top of the other along the spiral staircase, will allow the DNA to breathe much more than a string of six G-C pairs. Not only does the helix breathe, but, as in a badly constructed staircase, the rise and angle of the steps is not uniform.2 When an A-T letter pair is one step up from a G-C letter pair, the twist and rise of the G-C step are slightly different from those found when a G-C letter pair is above a C-G step. Just as you may trip running up a badly constructed staircase, the copying machinery is affected by the differing rise and twist of the steps in the helix. Such subtle differences in shape affect the accuracy with which the DNA itself is copied, and which mistakes are most likely to be repaired. These changes may be small, but they are not without effect. As the copying machine moves along a strand of the helix, it may pass along a GCAC or a GTGC. The physical structure of these sequences differs because of the way the letters “stack” on top of each other; some sequences are easier for an enzyme to copy than others. Variations in copying fidelity along the helix can be dramatic. Because of these shape changes, a particular mistake may be as much as 100,000 times more likely in one place in the genome than in another. Just as your jacket’s zipper may keep catching on a small bend in a metal tooth, some steps in a DNA helix simply are not copied as accurately as others. Copying may even stall at these places requiring enzymes to come in to cut and unwind the tangle. These differences also affect the accuracy of the error-correcting machinery. Different copying machinery and different error-correcting machines trip up in different places. Some mutations are so likely that they hardly seem random. Like accidents at a busy intersection without a stoplight, certain mutations keep happening. Some can be predicted, if you look closely. It is as if a local shopkeeper, watching the magic staircase on the plaza, could pick one particular step and say that, on the average, once a week or so, an apricot will slip in there next to the cobalt, instead of cobalt’s usual partner gold, at that very place. In her laboratory in New Jersey, molecular geneticist Lynn Ripley looks very closely, too.3 For example, she looked at mutations over many generations of a virus called T4, which kills bacteria. A section of the two-stranded sequence of T4 DNA that Ripley has studied with particular interest looks like this:
MUTATION IS NOT MONOTONOUS
37
TOP STRAND
ATGCTTCACGTTCGACGCTTCGACGGGAAGGGA TACGAAGTGCAAGCTGCGAAGCTGCCCTTCCCT BOTTOM STRAND
I’ll focus on just the top strand to describe what happened to this sequence. (Of course, because of the pairing rule, if I say that an A changed to a T on the top strand, that means that a T will be changed to an A on the bottom strand.) To a casual observer, it looks as if there were a variety of seemingly unrelated, random mutations in this region of the T4 DNA. In one mutated version of the T4 virus, the A in the 8th position had changed to a C. In another version, the 12th letter, T, had disappeared. In another mutant virus, a new A appeared between the 22nd and 23rd positions in the sequence. In yet another mutant, the G at position 26 was replaced by a T. No obvious pattern here; all very random, it seems. Yes, it seems random— except for one thing that Lynn Ripley noticed: She kept getting these same “random” mutations over and over again. If this was random, Ripley asked, why did she keep seeing the A in the 8th position mutate to a C, or the T in the 12th position disappear? Why did an A keep appearing between the 22nd and 23rd positions? Why did the G at position 26 get replaced by a T so much of the time? Why wasn’t it just as likely, she asked, for the C in the 7th position to mutate to a G or for an A to appear between the 20th and 21st positions? If this really was random mutation, why wasn’t it, well, . . . random? To visualize what might be happening, Lynn Ripley bent a representation of the sequence over in its middle, between the G at position 17 and the C at position 18, as illustrated in Figure 4-1. She could see that the letter pairs across two sides of the bent sequence matched correctly as if they were across from each other on two different strands of a DNA double helix, rather than both on one strand folded back on itself. When she bent the sequence of letters in the middle, the C at the fourth position was across from the G four positions from the other end. The T at the fifth position paired up with the A five positions from the other end; the T at the sixth position was paired with the A six positions from the other end. And so the pairing went up the center of the folded DNA sequence. It was almost perfect. The pairing of letters across this imaginary bent helix was almost perfect, but it was not exactly perfect. In addition to a small loop in the mid-
38
DARWIN IN THE GENOME
This T often is deleted
An A often is inserted here
This A often changes to a C
This G often changes to a T
Figure 4-1
Lynn Ripley’s explanation.
dle that was needed to allow the bending of the DNA like a hairpin and three letters that were unpaired at each end, there were a few mismatches. Lynn Ripley could see that the A at position 8 was mismatched—because there was a G, rather than the T needed to make an A-T pair, eight positions from the other end (at position 26). To allow the CGA starting at position 13 to match with a GCT 12 steps from the end (at positions 22 to 23), the T at position 12 would have to loop out so that the G would be across from a C rather than this T. Those few changes—a four-letter loop at the top, a single A to G mismatch, and a T looped out of the hairpin—were all that were needed to make a perfect hairpin; other than that, from position 4 counting from the beginning to position 4 counting from the end, there was a perfect little stretch of matched pairs. It became very clear to Lynn Ripley that her folded hairpin model could explain the mutations she kept seeing, over and over and over again. The single mismatch would be eliminated if the A in position 8 mutated to a C or the G at position 26 was replaced by a T because these mutations would create either a C-G or an A-T pair, “perfecting” the hairpin palindrome. If the T at position 12 disappeared, changing the GTTCG that begins at position 10 to GTCG, there would be no need for the imperfection of a looped-out T. GTCG would be perfectly matched with the CGAC that was across the little bent double helix, at positions 21 to 24. Or, if an A appeared between the 22nd and 23rd positions, replacing the CGAC at positions 21 to 24 with CGAAC around the bent helix, the T at position
MUTATION IS NOT MONOTONOUS
39
12 would now have a partner. This hairpin also explained why some of these mutations never happened at the same time. The A at position 8 might change to a C in the same mutant in which the T at position 12 disappeared, but she never saw an A appear between positions 22 and 23 in the same mutant in which the T at position 12 disappeared. What Lynn Ripley realized was that during copying, somehow, something like the little hairpin actually could form. Then the copying and errorcorrecting mechanisms that swirl around when each strand of a DNA helix is copied somehow would try to “fix” the doubled-over little helix, “fixing” the A-G mismatch at positions 8 and 26 to A-T or C-G, and excising the looped-out T at position 12 or giving it a partner by inserting an A between positions 22 and 23. Lynn Ripley looked at a little, imperfect hairpin loop and watched it be “corrected” (by mutation) to a more perfect helix. What had looked like random mutation was, upon very close inspection, not random at all. Or at least it wasn’t happening at a random place in the DNA sequence. The likelihood of each mutation depended upon its sequence context, that is, what other As or Ts or Cs or Gs were nearby, and in what order. Some mutations were orders of magnitude more likely than others. Looking at other similar sequences, Lynn Ripley could predict where they might mutate. A mutation that is predictable is not random. For this discussion, I have to separate two uses of the word random. It is clear, from work like Lynn Ripley’s, that mutations do not occur at random places in the DNA. However, they might be at random places with respect to the biological effects of the mutation. Whether or not they are random with respect to biological effects (for example, they did not appear to be random with respect to biological effects in the cone-snail toxins) is an important issue that I will return to soon. For now, I am just talking about where mutations happen in the DNA. Predictable mutations depend not only on the sequence of letters in the DNA, but also on which side of the double helix is being copied. Imagine that you are moving down the helical staircase, one railing on your left, the other on your right, just ahead of the machinery that is making a new copy of DNA. The “steps” that you are using to go down the center of the double helix are, of course, the letter pairs, A-T, T-A, G-C, and C-G. The railings of the staircase form the structural framework that links the two life-generating strings of letters. Each railing is a chain of the sugar deoxyribose, the D in DNA, and atoms that connect the sugars to each other. Because of the way the sugars are linked together, the railing has a pattern. The pattern is the same on both railings, but it has a direction. It runs down the railing on one side of the helix, and up the railing on the other
40
DARWIN IN THE GENOME
side of the helix. This pattern comes from the way the sugar backbones, the deoxyribose molecules that connect the letters, are hooked together. (When we talk about letters in DNA, like ATTCGG, we leave out the connecting sugars, but in real DNA, they are there between the letters, linking them to each other through hooks called phosphates.) If the sugar patterns on both of the double helix’s two railings went in the same direction, they might be described as parallel patterns; because the patterns go in opposite directions, they are called antiparallel. These antiparallel patterns complicate the copying of DNA. To follow along, you could draw two railings with antiparallel patterns, perhaps arrows running in opposite directions or something more creative. Now imagine separating the strands and copying each of them, as was illustrated in Figure 2.1. When DNA is copied, the strands separate, and a copy is made of each strand. Conceivably, this could have been done differently. It could have been that when DNA is copied, the result would be a completely new double helix, leaving the original double helix intact, but it doesn’t work that way. The strands separate and are copied, and each copy remains with the strand that it was copied from, the strand that was its template. The two original template strands, each with its new partner strand, leave each other, giving us two half-new DNA double helices. The enzyme machine that builds a new copy of DNA is designed to work in one direction, always building the pattern from left to right. But, here’s a problem for the DNA copying machinery: Because of the antiparallel railing patterns on the original helix, when the double helix opens up at one end for copying, the copying machinery is at the right end of the pattern on one strand of the helix and at the left end of the pattern on the other strand. The railing pattern on a new strand built from left to right starting at one end of the original helix will have the required antiparallel pattern, right to left on the original strand and left to right on its new partner. The problem is on the other side of the helix, where the helix railing pattern runs left to right. If it too is copied from the left using the DNA copying machine, which builds a copy with a left-to-right pattern, the new railing patterns will be parallel, not antiparallel as they need to be. This is how the problem is solved: To make a copy of the problem strand, the DNA copying enzyme machine slips a bit down the helix from the open end, then works its way back up, making a small piece of DNA called an Okazaki fragment after the husband and wife team that figured this out.4 Then the copying machinery slips even further down the open helix and, moving up again, makes another small piece of DNA until it reaches the Okazaki fragment it made just before. This new fragment of
MUTATION IS NOT MONOTONOUS
41
DNA is then attached to the first fragment, and so on down the helix. So, one strand is copied in this discontinuous, jumpy way. Because the machinery slips down a bit and copies up, the second strand of DNA can be copied with a backbone pattern that is antiparallel to the original, as it should be. There is still another level of complication in DNA copying, because a piece of DNA cannot be built by starting in thin air. Like a seamstress using a needle threader, the machinery that copies DNA begins by making a piece of RNA (similar to DNA, but with a different sugar, making it ribonucleic acid instead of deoxyribonucleic acid); the new piece of DNA then begins to grow, attached to the RNA needle threader. Each piece of RNA later has to be replaced with DNA before the new DNA strand that was made backward in little pieces is stitched together. When a little fragment of RNA/DNA is waiting to be processed, and the DNA copying and repair machinery is swirling around, pulling the DNA through in two directions, some unusual things can happen. Among other things, the little hairpin loops that Lynn Ripley envisioned have a chance to form. As you might expect, the types of errors made when DNA is copied tend to be different, depending upon whether the copy was made as one piece or in little pieces. On the leading strand, the strand that is copied straight through, the machinery is more likely to leave a mismatched letter as it rushes straight through. On the strand that was made in a discontinuous, jumpy way, the lagging strand, DNA sequences are likely to loop out and get lost.5 Thus many mutations are not random, at least with respect to their position in a DNA sequence. It clearly is not true that all mutations are equally likely, that any A in the DNA is just as likely to change to a G as any other. It is possible to assume that this is true and still do research if you study large numbers of sequences, in which the rare hot spots and cold spots may get buried in the noise of the average. But it really isn’t correct to make the assumption that mutations are random in the sense that they are equally likely at every spot along the DNA. Some changes are more likely than others. It is the other sense of random that I want to focus on now. This is the assumption that mutations are random with respect to their effect on the fitness of an organism. In other words, the assumption is that mutations that increase, that decrease, and that have absolutely no effect on fitness are all equally likely to happen. Once these mutations happen, natural selection will select the “fittest” from among these random mutations. Generally biologists figure that since the organism’s genome can’t predict what challenges its offspring will confront, mutations have to be completely random with respect to whether or not they might increase fitness.
42
DARWIN IN THE GENOME
The rare, lucky child gets a mutation that will help when a food source dies out or a new predator appears that begins to eat its sisters and brothers. The question I would like to raise here—and then return to with concrete examples, such as the cone-snail toxin, throughout the book—is whether mutation always is completely random with respect to its effect on the organism. In other words, can we really assume that how likely it is that a particular spot in the genome will mutate has no connection to how the mutation will affect the organism? In my experience, questioning this assumption invites controversy, or even the accusation that we are turning our back on Darwin, who proposed that selection acts on natural variations. Of course, Darwin died before he could learn about DNA, the genetic code, or even genes. But in the twentieth century, we learned much about genetic variation within a species and about mutations. We began to see what might underlie the variations in fitness that affect survival under natural selection. Once we learned about DNA, we saw that mistakes in copying generate mutations and so we’ve assumed that random mutation followed by selection underlies evolution.6 However, I have come to the conclusion that using our current knowledge to expand on Darwin’s insights does not require that all mutations be random with respect to their potential effects on biological function. We need to consider Darwin’s arguments in light of what we now know about intrinsic variations in mutation from spot to spot along the DNA. Much as the descendants of a bird experience selection based on variation in the shape of their beaks, descendants of an organism can experience selection based on variations in the intrinsic likelihood of mutation from spot to spot in their genome. Some types of changes are less risky than others. One single mutation may change the codon for one amino acid to that for a similar amino acid in a place in a protein where it makes absolutely no difference at all. Whether the amino acid at that spot is D or E, the protein still works just as well. But elsewhere in the protein there may be much less tolerance for change. There, changing even a single amino acid may damage or even destroy the protein’s ability to function. If that protein had an important role—say carrying oxygen to our tissues or starting the growth of the heart in an embryo—the small change might prove fatal, and not just for that one embryo; if this mutation were to happen very frequently in forming the sperm and eggs of those that carry this genome, it might lead to the end of the line for the genome itself. It is the survival of its children, and their children’s great-grandchildren, that carries the genome forward in time. If a lot of important information in a genome’s DNA tends to change, many children will not get that impor-
MUTATION IS NOT MONOTONOUS
43
tant information. If the children do not get that important information, they are likely to be at a disadvantage compared to the children of parents whose genomes did not store important information in such an insecure place. If a genome encodes copying machinery that tends to make risky changes, fewer of its children are likely to survive, compared to those of a genome with a copier that tends to make more creative changes. If fewer of a genome’s children survive because an important piece of information keeps getting lost, the genome won’t have as many descendants. Diversity and exploration—and, indeed, mutation—are needed, but it clearly is better if these creative forces do not tend to damage information that is essential for life. If important information is encoded in more stable parts of the DNA, generation after generation, this important information will have a better chance to survive and pass into the future. Genomes that survive natural selection will be those that tend to keep important information out of, for example, large disappearing DNA loops. The foibles of enzymes that repeatedly copy, move, and repair the sequence of letters in a DNA helix affect the evolution of that DNA sequence, and thus will fall under selective pressure. An example would be the alacrity with which enzymes correct the different “errors” in Lynn Ripley’s hairpin. Thus, not only individual mutations but the probability that certain classes of mutations may occur can become a substrate for natural selection. Because the chance of error at each step along the DNA staircase affects survival, the copying machinery itself, which affects the chance of mutation at each type of step in the helix, feels the pressure of natural selection. The locations, along a DNA helix, of accidental variations in the chance of a mutation can in this way be adjusted by selective pressure. Most differences in the probability of genetic change from position to position in the genome will be small, but some are large. In each genome, including each of ours, certain genetic changes are, simply put, orders of magnitude more likely to occur than other changes. Natural selection can act on intrinsic variations in the probability of distinct types of mutation along a DNA sequence just as it can act on all biological differences that have consequences for survival. The theme that occurs over and over again in evolution is that variations that may start out as accidental can be put to use. Charles Darwin pointed out that “the mind cannot possibly grasp the full meaning of the term of a hundred million years; it cannot add up and perceive the full effects of many slight variations, accumulated during an almost infinite number of generations.” He named “this principle, by which each slight variation, if useful, is preserved, . . . Natural Selection.”
44
DARWIN IN THE GENOME
Just as a breeder selects the prettiest roses, or cows that give the most milk, Darwin argued, Why . . . should nature fail in selecting variations useful, under changing conditions of life, to her living products? What limit can be put to this power, acting during long ages and rigidly scrutinising the whole constitution, structure, and habits of each creature,—favouring the good and rejecting the bad? I can see no limit to this power, in slowly and beautifully adapting each form to the most complex relations of life. The theory of natural selection, even if we looked no further than this, seems to me to be in itself probable.7 Now that we understand mutation at the molecular level, we should think again about this statement of Darwin’s: “Why should we doubt that variations in any way useful to beings, under their excessively complex relations of life, would be preserved, accumulated, and inherited?” As will be discussed in many chapters, the properties of a genome that lead to tiny, intrinsic variations in the probability of mutation along the DNA can be “in any way useful to beings” and thus become a substrate for natural selection.
5 What the Antibody Genes Tell Us The structure of the genotype is perhaps the most challenging remaining problem of evolutionary biology. —Ernst Mayr1
A
vertebrate genome first emerged on this planet about half a billion years ago. It led to orangutans, rabbits, chickens, sharks and whales, beavers, bears and bats, dinosaurs and dogs, and to us—to every animal with a backbone. Bacteria, viruses, fungi, and parasites already were here to greet us. To protect ourselves from the pathogens, we vertebrates emerged equipped with an innovative immune system. This chapter is devoted to the many levels of genomic creativity possessed by our immune systems. I am spending this much time on the vertebrate immune system because I deeply believe that for over 2 decades—ever since we discovered its “jumping genes” and its focused mutations—our immune system has been waving in our faces some important ideas about genome evolution. The vertebrate immune response demonstrates how a creative genome can be structured to handle the unexpected, how it can integrate a variety of mechanisms for focused genetic change into one effective system. Like the Lyme spirochete, our immune system can move pieces of DNA around the genome in a targeted, strategic manner, but with even more complex choreography. The DNA we inherit prepares us to resist repeated challenges by an extraordinary array of pathogens, but the information for making the anti45
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
46
DARWIN IN THE GENOME
bodies and other molecules needed to respond to these infections is contained in our DNA only implicitly. Our DNA does not protect us by “knowing” in advance the identity of every bacterium, virus, or parasite we will encounter; rather, it protects us by providing a strategic genomic infrastructure. This infrastructure facilitates efficient evolution of an effective response each time we are infected by a pathogen. It must be creative, for the pathogens that challenged our ancestors have continued to evolve. If we fail to respond, it is because the pathogen kills us too quickly, before our immune system can fight back, or because (as is the case with HIV) the pathogen damages the immune system itself, like a rogue gang that burns hospitals, ambulances, and pharmacies and kidnaps doctors and nurses. It doesn’t kill anyone directly, but it destroys the infrastructure that fights serious illness. Our bodies do have a few “canned” responses that evolved in our distant ancestors to handle common kinds of pathogens. For example, nine proteins in our blood, working together, can recognize and kill certain members of the gram-negative bacteria family. This family includes Escherichia coli, Salmonella, Vibrio cholerae, and other bacteria that can cause a wide range of diseases, from meningitis to pneumonia to food poisoning, especially when they build a capsule that protects them from our canned response against them. To get rid of bacteria that can resist our canned, predictable responses, we have an “adaptive” immune system. Our adaptive immune system is harder for a pathogen to evade because we keep changing, altering our DNA to create an effective response to the wide range of pathogens we encounter every day within each of our lifetimes. The discussion of the immune system in this chapter focuses on the antibody-producing cells, or B cells, which exhibit an impressive repertoire of unusual genetic activity. These B cells are not the only important immunesystem cells. T cells, which control much of the immune response and some of which can kill virus-infected cells, also play an essential role in protecting us, and employ many of the same genetic strategies used by B cells. Even the simplest virus is made of many protein molecules plus DNA or RNA. An antibody is just a protein molecule, and is much, much smaller than a virus. For an antibody molecule to protect us, it is not enough for it to grab onto a virus or other pathogen (unless the antibody molecule has managed to cover up the very part of a pathogen’s surface that the pathogen uses to stick to our cells). An antibody molecule sticking to a pathogen is just another protein molecule on the pathogen’s surface: For a virus, the antibody would be fur on its coat when viewed in an electron
WHAT THE ANTIBODY GENES TELL US
47
microscope; for bacteria, which are whole cells and are much larger than a virus, a molecule of antibody by itself would be nothing to be feared, or might possibly be food. Once an antibody has grabbed a pathogen, the antibody must trigger an effector response that ensures that the pathogen is destroyed. An antibody is a bit like a sentinel; it calls in the troops. For example, an antibody that has a pathogen in its grip can attach itself to a phagocytic cell, which can swallow them both and destroy the pathogen; or the pathogen-gripping antibody can trigger a whole complement of blood proteins that can poke holes in the pathogen’s surface. The design of antibodies, and therefore of the genes that encode them, has two somewhat opposed constraints, variation and conservation. On the one hand, the genome must provide a huge variety of antibodies in order to try out different ways of getting a good grip on an ever-changing landscape of pathogens. On the other hand, the genome must conserve those parts of the antibody molecules that are required for calling in the troops, whether that involves attaching to a phagocytic cell or fingering pathogens that use a capsule to hide themselves from aggressive blood proteins. A variable part of the antibody, then, is needed to grab whatever pathogen may come, to handle the unpredictable, while the conserved part is needed to carry out the more routine, yet essential, job of interacting with other cells and molecules to trigger one of the specific effector systems that our bodies have available to inactivate, destroy, and remove the pathogen. The tools for these two types of jobs are encoded separately in the DNA that we receive from our parents.2 The DNA that we received at conception did not come with intact antibody genes; the part of our genome responsible for our immune response came to us from our parents in pieces. It is as if our parents left the holiday gift in the box, unassembled, but with instructions, and went to bed. But don’t feel bad. Many vertebrates’ parents do that. We inherit three types of things in the genomic antibody gift box: a large variety of potential pathogen-binding pieces, a few regions of DNA that encode the pathogen-destruction information, and creative DNA-moving mechanisms that enable us to recombine these pieces. This genomic kit enables us to create and to explore a huge variety of antibody genes very quickly, while conserving the pathogen-destroying machinery. We have DNA encoding a large selection of potential pathogen-binding regions stored in our genomes, much as the Lyme spirochete stores a large selection of DNA that encodes different patches for its coat. Some of these variable, or V, regions will enable the antibody to get a grip on the mumps
48
DARWIN IN THE GENOME
virus; some will enable it to grab onto the measles virus; some will help it grab, we collectively hope, pathogens that neither we nor our ancestors have ever seen before. Our antibody genes are assembled, during each of our lives, from this palette of potential pathogen-binding V regions, together with a small set of C regions, conserved pieces of DNA containing information that encodes a short menu of ways to destroy pathogens. Before and soon after birth, temporarily protected by antibodies that our mother has given us through the placenta and in her milk, our immune system prepares for the inevitable onslaught of the many pathogens to come. Each B cell moves one of the bits of DNA that encode its hundreds of stored V regions next to a stretch of DNA that encodes C regions. (Actually, each B cell does this twice, as in each antibody molecule there are two kinds of protein chains, each with a variable region, that cooperate to hold onto a pathogen. The multiple possible combinations of these two chains generate further pathogen-binding diversity.) In order to create a fully functional B cell, the antibody gene-assembly machinery must find where the patches of DNA that encode V regions are stored in the genome. If we looked for these stored patches, we could use their special signal sequences to locate them. The signal is made up of two “words” that are separated by one or two turns of the double helix. One word has seven letters (e.g., CACAGTG), and the other has nine letters (e.g., ACAAAAACC), and they are separated by a spacer (like this message: scissorgnirtsrecapsrighthere). Such a message is easier to read in DNA because the letters in between the seven- and nine-letter words face the other side of the helix; the message is connected in space as if it were a tapestry attached to the railings on one side of a spiral staircase. Like tabs and inserts on paper dolls and their clothing, complementary DNA signals mark the borders of the C region where the moving V regions will land. Also separated by a spacer, these complementary signals are (written antiparallel, according to the A-T and G-C pairing rules, so that the first letter pairs with the last letter of the V signal in the previous paragraph) GGTTTTTGT and CACTGTG. As you can see, the two signals can pair with each other, a little like Lynn Ripley’s hairpin. Assembly instructions direct that the DNA be broken and joined between two words that are separated by one turn and two words that are separated by two turns of the double helix. In other words, if the V-region signal has about 12 letters (about one turn of the DNA helix) between its 7- and 9-letter words, the DNA it will become attached to has about 23 letters between its 7- and 9-letter words.
49
WHAT THE ANTIBODY GENES TELL US
V
J
V J
C
C
Figure 5-1 Moving the V region. A V (variable pathogen-binding) region is moved near the C (conserved effector) region. Variation is generated when one of a large selection of V regions is moved beside one of several J (joining) regions that sit beside the C region, identified by the signal sequences (indicated by stripes) to the right of the V region and to the left of the J regions.
This DNA moving (see Figure 5-1) creates our diverse repertoire of antibody molecules, each with a talent for grabbing onto something different. Like actors in a casting call, once the letters that encode the “auditioning” V regions have been moved from their storage place in the DNA we get from our parents to a place near the C-region DNA, they await the moment when their special talent is required. They await the pathogen that the antibody they encode is able to nab. Our genome has additional tricks for creating variation among our antibody proteins: In the very act of attaching one piece of DNA to another, it can vary the precise site of attachment. When the DNA encoding a V region and that encoding a C region are about to be attached, in the order V-C, a cut is made to the right of the V-region DNA and to the left of the C-region DNA. The precise position of this cut can vary in two ways, each of which generates additional diversity in the antibody. The first source of variation comes from a selection of one of several short patches of DNA, called joining, or J, regions, that are found in the genome at the left of the DNA that encodes the antibody C region. The new antibody DNA will look like this: VJ-C. V and J together encode what is considered the variable pathogen-binding site in the antibody protein. Therefore, depending upon which joining-region DNA is attached to the V-region DNA, the antibody will have different amino acids in its pathogen-binding region. There is yet an additional source of variation, which, again, focuses diversity right where it is needed, in the pathogen-binding site. This varia-
50
DARWIN IN THE GENOME
tion comes at the very site where the V and J regions are connected together. Suppose that the edge of the V region that is on the left side of the joint consists of the codon GAA, and the J region on the other edge, which will be on the right side of the joint, has the codon AGC. The assembly rule says that when the V-region and J-region DNA are attached, this connection can be made anywhere within or on either side of the three letters in each of these edge codons, as long as a total of three letters is chosen from the two codons at the edges of the V and J regions. As illustrated in Figure 5-2, without any mutation, depending upon where the connection is made when these same two pieces of DNA are joined, you could encode any one of four different amino acids at the site of the joint. It turns out that this very variable joint encodes a part of the antibody molecule that holds onto pathogens. It therefore creates a lot of extra variation; this extra variation further expands the range of pathogens that the antibody can bind beyond what is encoded explicitly in the genome through the repertoire of V regions. Of course, this extra variation at the joint is, in a way, also encoded in the genome, but not explicitly. Our immune system DNA allows us to create yet more variation at this important place in the pathogen-binding pocket of the antibody. Right at this place, where variation will increase the range of possible pathogens that can be grabbed, patches of extra letters can be added at the joint in the DNA that encodes one of the antibody protein chains. This extra diversity V
J
V GAA
C
AGC J V GAA J V GAC J V GGC J V AGC J
V J
C
Figure 5-2 Generation of variation in the pathogen-binding region. In addition to the availability of many V regions and several J regions, variation is generated through the ability to form one of four different codons at the site of the splice between the V and J regions, each one encoding a different amino acid.
WHAT THE ANTIBODY GENES TELL US
51
comes from small “diversity” regions, different from the V regions but also surrounded by the two words of seven and nine letters, with a spacer in between. One of these can be patched into the DNA near the C region along with the V region that is being moved. In this way, stretches of amino acids of various lengths can be added, which is a good thing, because different pathogens may have different “shoe sizes.” What’s more, while this variably sized diversity region is being attached, extra letters can also be added to the DNA by a creative enzyme that does not need to copy them from another strand. It can add a patch of “untemplated” letters; it appears to make things up. So, encoded in our genome is a set of mechanisms that generates a tremendous amount of variation in the pathogen-binding region of our future antibody genes. V regions are moved, splice sites are varied, additional diverse pieces are patched in, and extra letters are added—all generating diversity at the very spot in the antibody molecule that must be prepared to defend us against all of the pathogenic variety that can emerge in our environment at any time. The more varied our antibodies are right at that pathogen-binding spot, the better the chance that we will have an antibody that will be able to grab onto whatever pathogen infects us. In other words, the amount and location of genetic variation in the DNA that encodes our antibodies is matched to the functional requirements of the antibody protein. As big as the palette of V regions is, and as much as it may be enhanced by variation at the splice site, there is no guarantee that we have inherited a V region that can get a tight grip on any pathogen that might infect us; even if we have, a pathogen might mutate to elude the antibody’s grip. But if we have an antibody that can hold the pathogen even weakly, it is possible that a few well-placed changes in this antibody will allow the pathogenbinding V region to tighten the grip. Indeed, beyond the variation generated by the mechanisms I have described, our immune system generates still more variation in the pathogen-binding region of antibodies through hypermutation. Just as the cone snail’s toxin-encoding DNA seems to change more quickly than the rest of the snail genome, the part of our DNA that encodes the pathogen-binding part of our antibody genes can mutate much more quickly than the rest of our DNA. Mutation can tighten the antibody’s grip on a pathogen. The antibody genes begin to mutate after the V regions have moved into their new genomic context near the C regions. These mutations are specifically in the pathogen-binding variable region. Like Babe Ruth pointing to the precise place in the stands where he was about to hit a homer,3
52
DARWIN IN THE GENOME
our immune system has evolved the ability to focus mutations. There is information in our genome that directs mutation to the pathogen-binding region of antibody DNA. This information is found in the DNA near, but outside of, the DNA that encodes the antibody’s amino acid sequence. We know that this information is there because if you remove certain pieces of DNA that are near the rearranged variable region, you can stop the hypermutation. Or, if you replace the V region with a gene that normally does not hypermutate, say a piece of a hemoglobin gene, suddenly that gene will begin to hypermutate.4 This hypermutation is not spread evenly through the variable region, and that is a good thing. Mutation is more likely to happen in those very spots that create a pocket on the antibody’s surface that holds onto the pathogen, and it tends to keep away from the part of the antibody protein that makes up the scaffolding that holds the pathogen-binding pocket in place. One clue as to how mutation is focused may come from the way the antibody gene’s DNA encodes the amino acid S. There are six codons for the amino acid S. Therefore, wherever an S is needed in the protein encoded by the DNA, we would expect to find any one of these six. For example, AGC and TCA are considered synonyms because they both code for the amino acid S. Therefore, if natural selection favors an S in the protein, it really shouldn’t matter whether the S is encoded in the DNA by AGC or by TCA. But in the hypervariable part of the antibody gene, the amino acid S tends to be encoded using codons that begin with the letter A. In conserved places in the antibody gene, codons for S tend to begin with T. As our immune system tries out new pathogen-binding regions rapidly, the AGC S codons tend to change into codons that encode other amino acids, while the codons for S that begin with T tend to remain S codons. In the laboratory, you can change the odds that there will be a mutation in the V region just by changing the “synonymous” codon.5 As in cone-snail toxin DNA and Lyme spirochete DNA, codon synonyms behave as if they are not synonyms. Have we been missing something? There appears to be another message, “Make mistakes over here; we are trying new ways to combat a pathogen.” The antibodies, too, may phone us at 9 P.M. How hypermutation takes place is still a bit of a mystery, but it seems to be an active process, often focused, within the DNA encoding the V regions, on sequences with four letters in a special context. The first of the four letters can be G or A, the second is G, the third is T or C, and the fourth is A or T. As I write this, the evidence indicates that the process starts
WHAT THE ANTIBODY GENES TELL US
53
with a protein that targets a cut in the DNA to this spot.6 These cuts can be repaired, probably by a “sloppy” polymerase (a DNA-copying enzyme) or by patching in pieces of DNA from other V regions. People may use a sloppy polymerase; chickens and rabbits tend to use the patch. If we knock out certain proteins needed for “recombination”—the process that puts in the patch—chicken B cells seem to turn to the sloppy polymerase too.7 The details have not been worked out, but in fact a “sloppy” polymerase is turned on in immune-system cells.8 After all the gene moving and hypermutation needed in order to evolve, in real time, the right pathogen-binding region, our immune system is not done with its fancy footwork. Before I describe more genetic creativity, I want to mention, briefly, the cells in which all of this is taking place. The cutting and pasting of antibody genes takes place in the cells that become our B cells. Like a merchant displaying its wares, each B cell displays on its surface—its interface with the rest of the world—many molecules of a unique antibody. It can manufacture this antibody using the information in the DNA that it has patched together from V and C regions and perhaps some extra DNA in between. If this antibody bumps into and holds onto a pathogen, this plus a signal from other immune-system cells can trigger the B cell to reproduce itself, creating many, many more cellular factories that can manufacture the same selected antibody. This increase in the number of factories that can make the selected antibody vastly increases the number of antibody molecules available to bind to this specific pathogen. The cells that serve as these extra factories survive for a long time, providing immunologic memory of the attack (or memory of a vaccine that looks like the pathogen) and leaving us well stocked and prepared to manufacture the right antibody, right away, to combat another attack by the same pathogen. Thus we are immune to this pathogen. Once there is enough of the appropriate antibody, it must direct an effective attack on the pathogen, and to do this it must have options. Pathogens are not predictable, and they don’t come with instructions. One pathogen may try to establish a foothold in the blood, another in our guts. One may be a virus, which commandeers the protein-making machinery of our own cells. Another may be a parasite, which attaches to us and feeds upon us. An antibody may have to travel out into tears to protect the surface of an eye, or circulate in the blood and attach the pathogen to phagocytic cells; it may need to cross the placenta or trigger the release of histamine. But before an antibody can direct any of these attacks, it has to grab onto the pathogen. And in order to do the appropriate thing once it does grab onto
54
DARWIN IN THE GENOME
the pathogen, it must be attached to the appropriate effector region. As you might expect, in order to get the right binding site to the right effector region, the immune system undergoes still more genetic gymnastics. The DNA that encodes the variable region that just has evolved by hypermutation to grab the pathogen from the B-cell surface can move yet again in the genome of our B cells and so become attached to DNA encoding any one of seven other distinct effector domains that can dispose of the pathogen. As the infection proceeds, our immune system will switch the pathogengrabbing variable region from the first C region, which holds it onto the B-cell surface, to a different effector region. To move the pathogen-grabbing region to the new effector region, in precise genetic surgery, the DNA must be cut. The cut end near the pathogen-grabbing region must be joined to a cut end near the effector region that is appropriate to the task at hand. When we are under attack by parasites, for example, the ε effector region is likely to be most useful. Antibodies that protect us against Staphylococcus, to use another example, can go into our tears if they are attached to effector domain α, or they can go into our bloodstream if they are attached to effector domain γ. Antibody-producing cells cut their DNA in specific locations in response to a signal from other cells in the immune system. These other cells let the antibody-producing cell know which type of response is needed, and thus which effector region should be attached to the antibody. Because it is a bit startling to think about making a regulated cut in the DNA of an active gene, I want to take a moment to describe how this might happen. The signal that triggers the DNA cut that switches the pathogenbinding region to a new effector region seems to start out like a typical signal from one cell in our bodies to another: A signaling protein, this one called a cytokine, travels from a cell that senses the type of invader to the antibody-producing cells. This cytokine attaches to the surface of the antibody-producing cell and, through a relay of protein signals that reach from the outside of the B cell into its chromosomes, turns on specific genes. “Turning on” a gene means copying the gene into RNA, and that is what happens with the antibody genes. Like many of our genes, the antibody genes have introns, additional letters in the DNA between the letters that encode the V region of the protein and the letters that encode each effector region. Introns are copied into RNA, but they are spliced out of the RNA before it gets to the protein factories. When antibody genes are about to be cut in order to switch effector regions, something unusual happens. The machinery that copies information into RNA does not copy the proteinencoding part of the V-region DNA; rather, it starts copying in the intron,
WHAT THE ANTIBODY GENES TELL US
55
just after the letters that encode the pathogen-grabbing region, so that it copies through the place where the “switch region” DNA will be cut.9 With one strand of the DNA holding onto the RNA copy, the other DNA strand in the double helix can “loop out.” This switch-region loop will be full of Gs. If it loops out and is left to its own devices, this G-rich strand can form a very unusual structure, with 4Gs in a plane instead of a helix, with these planes stacked on top of each other. As there is an enzyme that can cut DNA next to these unusual G structures, it may be that this enzyme makes the first focused cut in DNA10 although the DNA in this region also is rich in palindromes, which may focus the cut. Eight of our effector regions are near such switch-region DNA. When an RNA copy is made of the switch region appropriate to the pathogen at hand, that switch region DNA is cut. The V region is attached to the appropriate effector region by cutting the DNA between the switch region to the right of the pathogen-grabbing region and the switch region to the left of the appropriate effector region out of the genome, leaving the pathogengrabbing region attached to an effector region that once was many letters further away. When the new antibody protein is made, the variable region that evolved the ability to grab the pathogen now is attached to a new effector region in one seamless protein chain. It is clear that our immune system has a large repertoire of unusual genetic tools at hand. It can recognize specific sites in the DNA where it can cut, move, and attach pieces of DNA. It can focus variation on a small region of the DNA, and, within that region, on specific spots. These spots produce variation exactly where it is needed, in the pathogen-binding region of the antibody. Then our immune system cuts active genes to generate an appropriate pathogen-destroying effector response. Our immune system has integrated a range of creative tools for genetic exploration that generates effective responses to likely challenges. It faces, repeatedly, the need to hold onto any one of a huge number of potential types of pathogens, and then to dispose of them in one of a few well-defined ways. The lesson that our immune system teaches us, dramatically, is that mutation can be focused at an appropriate site within a gene in a way that is regulated by signals in the cell’s environment; whole regions of DNA can be moved around in a regulated, evolutionarily favored way; and DNA can encode information that attracts mechanisms that cut and paste together pieces, or modules, of DNA at specific places in the genome, creating hot spots of genetic variation. Hypermutability can be intrinsic to a specific DNA sequence placed in a specific genome sequence context.
56
DARWIN IN THE GENOME
While antibodies are selected one at a time for their ability to bind to a specific pathogen, the infrastructure that creates antibody genes was selected on a higher level for its ability to focus genetic mutation in the appropriate spots. What has emerged from natural selection is a genetic infrastructure that rapidly creates new binding sites, while conserving what it takes to be an antibody. Most important of all, the antibody DNA demonstrates that genetic hot spots of variation can evolve in appropriate locations, locations that are well matched to the actions required of the protein that is encoded by the gene. In the immune response, variation is focused at the pathogen-binding region, the region within a gene where variation is most likely to create useful new antibodies. In evolution, only some variations are favored by natural selection. So, too, in the immune system, not every rearranged V region will be selected for by binding to a pathogen. However, it is an effective strategy for a genetic infrastructure to provide a diverse selection of V regions—all variations on a theme, with diverse pockets available to bind diverse pathogens—along with a menu of options that enable the immune system to get rid of different classes of pathogens. We are ready. This infrastructure gives us better odds, compared to random variation, that we will be able to protect ourselves against each pathogenic challenge that confronts us. Like the snails and spirochetes, our antibodies illustrate that biochemical tools can evolve that can make genetic variation, i.e., mutation, more efficient than random mutation in evolving useful new functions. With all of these genetic tools lying around, available for focusing variation, it seems reasonable to suppose that they would have been picked up and used for other creative explorations in the genome. There is no reason to assume that the use of these tools is limited to battles between pathogens and their prey. Indeed, some of the proteins that move the antibody genes and generate variation are turned on in our germ cells as well, and, interestingly, in our brains. Perhaps the dramatic viper/snail, pathogen/ prey battles are simply the ones that have drawn the attention of our research programs, just as the fires and murders draw the attention of the TV local news while the day-to-day work of educating children generally goes unreported. I have been told that mutation must remain random with respect to function because a genome cannot predict the future. Can you and I predict the future? Even though we can’t exactly predict the future, our actions are not completely random. We go about our lives each day, making informed guesses about what will happen next. We can do this because
WHAT THE ANTIBODY GENES TELL US
57
events in the world around us are not completely random. Certain types of challenges tend to confront us again and again. We discover, starting at a young age, that some types of responses to those recurrent challenges tend to work better than others. A genome can “see” into the future the same way we do, on the basis of what it has experienced, and survived, in the past. But wait a second— aren’t we different from genomes? We are conscious; we can learn; we are not strings of letters. Let’s meet again for this discussion toward the end of the book. In the meantime, let’s hold this thought: We excel when our brains capture information and learn, efficiently, from experience. Perhaps our genome has been structured, by natural selection, to be efficient at evolving.
This page intentionally left blank.
6 Slippery DNA and Tuning Knobs In a scientific context, the word spontaneous is meaningless. —Barbara E. Wright1
E
ven if you were a highly evolved DNA copying machine, it would be easy for you to lose your place and miscount when you kept running across something like TTTTTTTTTTTTT, at 80 to 500 letters per second, time after time, generation after generation. In some bacteria, each of these repetitive sequences increases or decreases in length about once every 10,000 times the DNA is copied; that means that among the descendants of an individual bacterium, which can divide in two every half hour, each of these sequences can slip many times each day. Each slip makes one sibling unlike another. Each slip generates diversity. Copying and repair machinery also can miscount repetitive sequences such as CAGGGCAGGGCAGGGCAGGGCAGGG. The new strand can slip and partner up with the wrong CAGGG on the other strand. This kind of slip results in copies of the gene that have more or fewer CAGGGs in a row. One of the most unstable places found so far in any DNA is a part of the mouse genome that can have from 200 to over 1000 repeats of CAGGG.2 These CAGGG strings can change in length by more than 1000 letters between parent and child mouse. As the cells of an individual mouse divide to form the mouse’s tissues, this string can slip so often that the mouse’s DNA differs from tissue to tissue. 59
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
60
DARWIN IN THE GENOME
Remember the badly built staircase, with each rise and tilt being a little different, depending upon the DNA sequence. When you run up a flight of stairs, you are not thinking about raising your leg the right amount to reach the next stair; your body adjusts your steps subconsciously based on the first stair or two. So, if, in the middle of a staircase, a carpenter builds an individual stair, or a run of several stairs, with a different rise, angle, and depth, someone is likely to stumble. When a tilted DNA sequence repeats stair after stair, a whole stretch of DNA can have a very unusual shape, throwing the rapidly moving copying machine off of its game. Some repetitive sequences even can fold back on themselves when the two DNA strands come apart and are being copied. A string of repeated CTGs tends to loop out into a hairpin in the DNA strand template that is copied backward. When that happens, if the hairpin is long enough and stable enough, all the CTGs in the loop will be missed as the copying machinery moves by, and the new generation will be missing copies of all the CTGs that were in the loop. On the other hand, if a loop bulges out in the new strand of DNA that is being built, then the copying apparatus may not know that these CTGs are there; and so some of the CTGs in the original template strand may be copied again. These progeny get extra CTGs.3 When short strings of repeats are miscounted, the mismatch repair system sometimes can fix the mistake, but if the repeats grow too long, the jumps become too big to be fixed by mismatch repair. As the repeats grow, the chance that they will get even longer becomes from 3 to 175 times greater than the chance that they will get shorter, and the size of each change increases. Once the repeat is long enough, it can increase by as many as 1000 letters in a single jump. Long CGG and CTG triplet repeat sequences keep bending in the same direction, making loops that turn back on themselves every 81 letters (or 27 triplets) to form toroids, which are structures that look like a helix wrapped around a donut. When the DNA strands come apart for copying, a string of GAAs can form triplexes, in which three strands line up: one of them the string of GAAs, one its original partner, a string of TTCs (read backward according to the antiparallel rule), and the third another string of GAAs copied from the partner. These three-strand structures are held together with their own new pairing rules: Two of the strands have an A or a G, the third strand a T or a C. When one stable triplex structure of this type was followed, researchers found it slipped so often that 96 percent of the children inherited a different number of repeats.
SLIPPERY DNA AND TUNING KNOBS
61
Some DNA sequences, such as strings of CGG, can even form fairly stable tetraplexes when they are opened for copying. In a tetraplex, two hairpins can form on the same strand and then stick to each other, breaking the usual A-T and G-C pairing rules. For example, two Gs from each step of one hairpin loop can stick to two Gs from each step of the neighboring hairpin loop.4 If the repeated G-rich sequence is long enough, many tetraplexes will form along the DNA chain and stack together, forming hollow stems or cylinders. Just when you think you have gotten to know it, DNA can begin to remind you of a seemingly conservative neighbor who one day invites you to a funky gallery on the fringes of SoHo to show you her mazelike floorto-ceiling-to-walls installation of intertwined, multicolored string artforms. Especially in the midst of the action, when its strands come apart, and it is being copied in two directions at once, DNA can form unusual structures and behave in unexpected ways. Once the normal copying structure is disrupted, DNA can become unstable, creating a gap—a slip between cup and lip—between this generation and the next. When the copying machinery slams into unusual structures, such as tetraplex cylinders, it stalls. Okazaki fragments, the bits of the new partner to the strand of the helix that is copied discontinuously, can fold back and not be trimmed properly. A repair system may cut the DNA at the stall, copy some sequence from another strand, bring in an error-prone enzyme that may stick in the wrong letters—and the result is a mutation “hot spot.” When the DNA copying machinery is idling and cannot move forward, this increases the likelihood that it will slip again and slip again, copying the same repeat over and over again at the edge of the block and increasing the length of a large repeated DNA sequence. If the tetraplex is built of repeats of CGGCGGCGGCGG, then the stalled DNA copying machinery may keep copying CGGCGGCGGCGG, inserting many extra copies before it is able to move on. These strings may increase in length many times within the space of only a few cell divisions. CGGCGGCGGCGG’s usual partner strand, with repeats of CCGCCGCCGCCG (again read backward), does not form these stable tetraplexes, so the copying machinery doesn’t stall when it copies that strand; thus the new double helix that results from copying the other strand keeps the original number of repeats. This means that one of the two children gets the expanded string and the other doesn’t, creating diverse children. A very common slip occurs at repeats of triplets, such as CAG. Because the slip is exactly three letters, which, in the case of CAG, encode the amino acid glutamine (Q), the children inherit genes that code for proteins that are identical except for increased or decreased numbers of glut-
62
DARWIN IN THE GENOME
amines. Too many Qs in a row can make the proteins sticky and can do some damage. Woody Guthrie5 was born in a small frontier town in Oklahoma in 1912. The family was poor, his sister was killed when their house burned down, and his mother was stricken with a mysterious illness and institutionalized. Woody set to wandering, at the time of the Great Depression and the dust bowl, meeting many other poor and hungry migrants along the way. He spoke out—sung out, actually—for the cause of those who were oppressed. He hosted a radio show, and he wrote children’s songs and songs such as “Union Maid” and what has been called America’s second national anthem, “This Land Is Your Land.” Constantly in motion, outspoken about what he believed in, standing up and fighting for the unfortunate, Woody Guthrie seemed like someone whom nothing could keep down. But something did knock Woody Guthrie down. He began to move and behave erratically. Some people thought he was suffering from schizophrenia or from alcoholism. By 1954, he had checked himself into a hospital, but he continued to deteriorate. In 1967, he died. It turned out that Woody Guthrie, like his mother before him, had Huntington’s disease, an inherited disorder. In 1993, 26 years after Woody Guthrie’s death, through efforts spearheaded by Nancy Wexler, the daughter of a Huntington’s disease victim, we learned that the tragedy of Huntington’s disease is a legacy of the slipping of DNA when it is copied.6 An expanded repeat of three letters, CAGCAGCAGCAGCAG, killed Woody Guthrie and, later, two of his eight children. The first symptoms of Huntington’s disease are subtle and easy to ignore—just some slight clumsiness or mild absentmindedness. But in a family that knows that it carries this disease, every dropped key and forgotten phone number can be cause for silent panic: Is it starting? Is it happening to me? Clumsiness and unsteadiness in walking lead to random involuntary movements, such as jerks, twitches, and flailing arms and legs. Speech becomes more and more slurred, and eventually the victim cannot walk, talk, or stand. What began as mild irritability, forgetfulness, or depression gives way to disorientation and dementia; and death in about 10 to 15 years.7 Although there have been Huntington’s patients ranging in age from 2 to 80, the disease usually strikes in middle age, after its victim has had children. So patients and their families have to worry that the disease is lurking like a time bomb in the DNA of their young children. Huntington’s is a dominant genetic disorder, which means that there is a 50/50 chance that a child has inherited it from an affected parent.
SLIPPERY DNA AND TUNING KNOBS
63
As if that’s not cruel enough, the CAGs tend to keep expanding; so a child may well have inherited an even longer repeat, and thus the disease may strike this child at a younger age and be even more aggressive. In other words, Huntington’s disease exhibits anticipation: The severity increases and the age of onset decreases in successive generations. As the length of the CAG repeat grows, so also do the size of the successive expansions and the likelihood of yet another jump in size. Expansion to more than 90 letters can occur in Huntington’s disease patients, and this makes a string of more than 30 Qs in a row in the protein. Families without any affected relatives have not necessarily dodged this bullet. A bad jump can bring Huntington’s into a new family, as the length of the string of CAGs grows beyond the outer boundaries of unaffected. In the gene that slips in Huntington’s disease, people with 28 repeats are fine; their DNA doesn’t tend to slip. DNA with 38 repeats often slips. DNA with 60 repeats almost always slips. A string of 30 to 38 repeats is just at the boundary and is common; it is found in an estimated 1 in 50 people in the United States. In unaffected families, a string of CAG repeats is often interrupted by a CAA. If that CAA mutates to a CAG, there can suddenly be great instability. This expanded gene seems to enter new families after something slips when the father is making a sperm. Scientists examined CAG repeats in over 3500 sperm8 from men with repeats ranging from 37 to 62 CAGs, and found that there were slips in 82 percent of the sperm. Both the number of slips and the size of the jump increased as the number of repeats increased. For men with at least 50 repeats, 98 percent of their sperm had slipped. Individual sperm from each individual had different lengths of the expanded repeat. Right now, Huntington’s disease is rare—but not rare enough, of course— in the United States; about 25,000 people suffer from it. But in one small village on Lake Maracaibo in Venezuela, a large fraction of the people you are likely to meet will be struggling with the effects of Huntington’s disease. They are the descendants of a woman whose DNA had slipped, perhaps in the one of her father’s sperm that had fertilized her mother’s egg. Huntington’s disease isn’t the only serious disorder linked to expanding repeats in DNA. In fact, there are at least 14 such diseases that strike people.9 Like Huntington’s, eight of these other disorders resulting from slippery DNA involve expansion of the triplet CAG in different genes, thus encoding strings of Q repeats in distinct proteins. People born without the time bomb may have 5 to 30 CAGs, and thus 5 to 30 Qs, in the protein. But when the string of CAGs slips to make a protein with 40 to 100 Qs in a row, the proteins begin to stick together in long-lasting clumps and interfere with the normal on/off bumping and parting of proteins.
64
DARWIN IN THE GENOME
In four other time-bomb disorders, triplet repeats expanded in regions of DNA that are not actually translated into protein, but are copied into RNA. In myotonic dystrophy, for example, a string of repeating CTG triplets expands from a normal range of 5 to 40 repeats to a disease-causing range of 50 to 3000 and beyond, jamming up RNA handling systems and interfering with the translation of information in the RNA into proteins.10 We are not the only ones with slippery DNA. Bacteria too have repetitive DNA sequences, such as CAATCAATCAATCAAT, that often slip, growing and shrinking in size by multiples of the size of the repeat, in this case four letters. Because the genetic code is read in groups of three letters, a slip of four letters throws the message out of frame, mgitun in telligibletoth er ead er. In fact, repeated shortening and lengthening of the DNA by four letters disrupts, then restores, then disrupts again the function of many of the bacteria’s genes. But, far from causing trouble for the bacteria, this bacterial slippery DNA causes problems, once again, for us.11 How does a bacterium like Haemophilus influenzae survive in our nose and throat, which are constantly bathed in protective antibodies? H. influenzae keeps changing its coat by turning genes off and on by slipping at repeats. Not only do these ever-changing coats keep the bacteria a step ahead of the immune response, but they even may find a coat that doesn’t trigger the immune response. These rapid changes enable the bacteria to avoid antibodies, but that is not the end of the trouble that they cause for us. H. influenzae has repeats of CAAT and Neisseria meningitides has repeats of CTCTT that both slip and change their surfaces. Not only does this help them to evade antibodies, but this changing of their coats may change the repertoire of tissues that the bacteria can invade during an infection. To invade, bacteria have to stick to us. They have evolved many specific surface structures that permit them to stick firmly to our “inside” skin, the skin that lines our gut, nose, ears, throat, etc. Because bacterial coats are stitched together from many fabrics, changes in several genes that encode the proteins that put together the coat allow the bacteria to mix and match patterns in different combinations. This combinatorial exploration of surfaces allows them to explore different ways of attaching to us. Suppose, for example, that you had a palette of five possible decorations for a wall, each of which is available if a specific “decoration” gene is on, and is not available if that gene is off. Say one of the genes encoded information that created a circle with a hook to attach it to the wall, another created a triangle, another poured red paint into a can for use in painting the shape, another poured yellow paint, and the last encoded the
SLIPPERY DNA AND TUNING KNOBS
65
ability to create ribbons and tie them to the wall. So, depending upon which genes are on, your wall could be decorated with a circle painted yellow with ribbons, red triangles streaming with red ribbons, orange ribbons, and so forth. In the same way, by turning genes on and off, bacteria “decorate” their surfaces with different coats that allow them to hide from our immune system and to stick to different places within us. The bacteria get even more sophisticated than turning genes either on or off. DNA slips also can vary the amount that the genes are used. It is as if, when you are decorating the wall, you were to tinker with the amounts of red and yellow paint that you poured into the can. Rather than one evenly mixed orange color, you would have different gradations of orange depending upon the relative amounts of red and yellow that you used. In order to use the information in a gene, we first have to find the gene. After all, DNA is just a very long string made up of four different letters. DNA encodes more information than just the order of amino acids in the protein. There is information in the DNA that is copied into RNA that does not become protein, but that helps the RNA get translated and determines how long it lasts before the message is destroyed. There is other information in DNA that is not even copied into RNA. For example, there is information that tells the RNA copying machinery where a gene begins. The recognition signal for the start of a gene involves two specific groups of letters, which are found about 10 and 35 letters before the place where copying should begin. If these two groups of letters are the correct distance apart, this points the copying machinery to a gene. What is the right distance apart? Actually, there is a little flexibility. There are optimal distances, but there also are longer and shorter distances that can work, although maybe not as well. Longer and shorter distances change how well the copying machinery fits, how efficiently it starts, and how often the information in the gene is turned into a protein. For example, in the DNA that encodes the H. influenzae surface structures that stick to us, a spacing of 16 letters results in the highest level of copying. If the spacing is just two letters shorter, 14 letters, the gene will not be copied. For N. meningitidis, spacers with a run of 11, 10, or 9 Gs between the telltale gene-marking groups result in high levels, medium levels, or no creation of messenger RNA, respectively.12 The length of a repeat not only turns a gene on and off as it slips from generation to generation, but also changes how sensitive the gene is to being turned on and off by molecules in the environment. In one example, how sensitive a gene was to being turned on and off by a specific compound in the environment depended on the length of a run of TTTTs.13
66
DARWIN IN THE GENOME
Slips that keep turning different sets of genes on and off generate diversity, the molecular equivalent of changing the availability of decorations for a wall. These slips generate a repertoire of variants from which the bacterial cells that are the fittest can be selected. In this case, “fittest” can mean the best at attacking us, avoiding our immune system, and sticking to and entering our tissues. The infection process is a dynamic one. Variants that are fit at one place, say when they enter our throats, may not be well adapted for life in our bloodstream. N. meningitides builds a capsule around itself, which allows it to resist being killed by our immune system. But this protective capsule poses a dilemma for the bacteria, as it also interferes with its ability to attach to our throat. Most meningococci carried in people’s upper respiratory tract are missing the capsule. However, during one outbreak of meningococcal disease studied closely by researchers, capsular forms predominated at the sites of invasion of tissues.14 An insertion or a deletion of one C in a string of seven Cs, which happens about once in every 1000 bacterial cell divisions (possibly occurring a few times a day) stops the copying of the gene for the capsule. When the disease broke out and the immune system was called into action, the DNA had slipped back; isolates had exactly seven Cs in that string, and the deadly bacteria were encapsulated and protected from attack. These bacteria are able to adapt their genes to generate variation depending upon their location inside us—a tremendous advantage. It is as if you could vary the bottom half of your body to optimize your experiences during a day in a lakeside village: The bottom half of your body would become wheels to get you down the road, then would be flippers when you entered the lake, and finally would turn into a lawn chair at the end of the day when you wanted to relax and watch the sun set over the water. While short repeats, such as the seven Cs, seem to be mainly involved in turning genes on and off, bacteria also have longer repeats that can adjust properties of their proteins. For example, in a Staphylococcus aureus molecule that sticks to us, the variability of the numbers of repeats of the amino acid pair QS works like an adjustable stalk, almost like a jack raising a car, to allow the part of the protein that sticks to us to get up high enough above the bacterial surface.15 Because of their strategically slippery DNA, bacteria can keep generating diversity, allowing them to find the right approach to challenge our defenses and to move through the different microenvironments of our body. These slippery DNA sequences are found in many bacterial genes, but are absent from many others. And these groups seem to be different kinds
SLIPPERY DNA AND TUNING KNOBS
67
of genes. Genes in one group, which have been called contingency genes,16 demand flexibility—for exploration, to respond to challenging circumstances, and to adjust to the environment. Different combinations of these genes are needed if the bacteria are to bind to our different tissues. These contingency genes appear to be different from housekeeping genes, which are needed to run the more predictable everyday business of turning starch into sugars and copying the information in DNA into proteins. The ability to adjust their surfaces to explore new territory, to vary their coats to avoid capture by the immune system, and to enter and colonize new environments—all by selectively changing the most appropriate subset of their genes at a high rate, without waiting for mutation to damage the rest of the DNA—is an effective strategy for bacterial survival and evolution. But is it only bacteria that have captured the value of slippery DNA? Walter Schaffner,17 Director of the Institute of Molecular Biology at the University of Zurich, Switzerland, suggests that slippery DNA is important in other species as well, especially in genes that affect the extent that other genes are turned on and off. Evolution depends not just on the creation of new genes, but on changes in the activity of the genes we already have. One example of the importance of slippery DNA to animals comes from fruit flies. Fruit flies may be little, but they are much more similar to us than bacteria are. Like us, fruit flies come from fertilized eggs and grow muscles, a gut, and a brain. It may be hard to recognize a sleepy fruit fly, but, again like us, fruit flies have biological clocks, 24-hour wake/sleep cycles. These cycles are managed by genes that are similar to the genes that manage our own wake/sleep cycles. Fruit flies do not maintain their body temperature, however, and so they have an additional challenge: to keep their biological clock running accurately at different temperatures. In one of the clock genes of one kind of fruit fly, a string of the amino acids TG changes in length. At warmer temperatures, fitting its Mediterranean lifestyle, a variant with 17 TGs in a row can keep its clock running at close to 24 hours.18 At colder temperatures, this variant’s clock runs a bit too fast. Another variant, with 20 TGs, has a clock that works about the same at both temperatures; it runs a little fast, but it’s more accurate than its Mediterranean cousin’s clock in northern Europe, where this variant lives. These effects are marginal, but it is possible that a slippery gene coding for the amino acid pair TG gives a slight advantage to the fruit fly. Ed Trifonov,19 who runs the Genome Diversity Center in Haifa, Israel, and David King20 of the University of Illinois have suggested that slippery sequences like TTTTTTT or CTGCTGCTG would be useful to our
68
DARWIN IN THE GENOME
genomes too. Trifonov and King independently began to use the metaphor that, because these sequences change in length so easily, they can act as “tuning knobs.” The activity of genes might change frequently between generations, increasing and decreasing how often the information in the gene is copied into RNA as the number of repeats increases and decreases when the enzyme that copies DNA to DNA slips. Some slips will hurt, some slips will help, but each slip ensures that among the human family there is variation in the activity of certain genes. It generates diversity. These slips may even be relevant to another observation of Darwin’s, anticipation. “I could give a good many cases of variations (taking the word in the largest sense) which have supervened at an earlier age in the child than in the parent.”21 Perhaps anticipation allows a genome to try something out in adults first, then move it to younger and younger ages if it is a successful variation. We know that anticipation can occur by expansion of triplet repeats, as we saw with Huntington’s disease. Finding out whether the expansion of repeats might be among the mechanisms responsible for the anticipation that Darwin observed will require the sequences of many more genomes. We do not know yet if this is true. If asked to bet, I would guess that it is. The correct balance between fidelity to the text (careful copying of the genome) and variation (genetic exploration) emerges though natural selection. The correct balance does not have to evolve with an even, genomewide, probability of change; it may be better to have different rates of change in different places. In some places, such as in contingency genes in bacteria, changes create variation that is adaptive. Genetic variation at each position in a genome becomes a conversation, across generations, between DNA and the enzymes that repair, copy, and move it. And what about us? Repetitive sequences occur in thousands of places throughout our genome. Strings of 1 to 5 letters may be repeated in tracts that are from 10 to 1000 letters long. Elsewhere, strings of 5 to 100 letters can be repeated in tracts that are from 400 to 30,000 letters long; these vary so much between people that they are used as genetic “fingerprints”22 in paternity suits and criminal investigations. While long strings of CAG cause tragedy, as in Huntington’s disease, shorter strings of CAG are found in many genes. Perhaps variation in the string of Qs in the Huntington gene is good until the string gets too long. Many proteins with strings of Qs can turn on the copying of certain genes and block the copying of others. For example, there is a protein called the androgen receptor that detects the presence of testosterone and turns certain genes on or off when testosterone is present. The gene for the androgen receptor
SLIPPERY DNA AND TUNING KNOBS
69
has a string of CAGs, and so the androgen receptor protein has a string of Qs. When this string gets longer, the androgen receptor is less effective at turning on some specific genes. The glucocorticoid receptor and many other proteins that are part of control systems that turn genes on and off also contain tracts of Q-rich regions. So far, we know that Q strings are present in more than 30 proteins involved in regulating the extent to which the information in particular pieces of DNA is transcribed into a messenger RNA.23 Strings of Qs often are involved in binding proteins to each other; in fact, they have been called a protein “zipper,” linking two proteins together. It is very likely that the strings of Qs in the tragedy of diseases such as Huntington’s are just too long and too sticky, but shorter, variable strings of Qs probably are very useful in adjusting binding. Strings of CAGs in parts of the DNA that are not copied into protein certainly could have an adjustable, tuning-knob-like effect. When the gene is transcribed into RNA in preparation for making a protein, there can be a dramatic decrease in the amount of that RNA that gets translated into protein as the string gets longer. Like the bacteria, we have probably captured value from slippery DNA. Our inheritable traits are not so discrete as “blue” eyes or “brown” eyes. We see subtle differences all around us that involve gradations, say of height or of facial features. It is likely that some genetic mechanism is generating a subtle range of diversity frequently. One way to generate diversity would be with a combinatorial mix of levels of the proteins encoded by different genes. Rather than being just a matter of turning genes on and off, variation can come from adjusting the levels of expression of the genes, like different tones of orange paint. Because of the interconnectedness of the systems that regulate our genes, mutations that alter the expression of one gene can change the expression of genes that have not mutated.24 This type of variation is bound to affect our fitness. It certainly does so in very special circumstances. The number of repeats next to an enzyme that is targeted in cancer treatment, for example, affects the amount of the target protein that the cancer cells contain, and recently was found to predict whether the treatment would work.25 Now that our genome is available, we will be able to connect these slippery DNA regions to their tuning-knob role, if they have one. But we will be able to do this only if we look into our genome with some respect, seeking to learn. We will not find them if we dismiss “boring” repetitive sequences as “junk DNA.”
This page intentionally left blank.
7 Everyone Has Something to Teach Us We are a small, although important, part of this wonderful world, the development of which is based on deeply anchored interdependencies. —Werner Arber1
W
hether it is poking and prodding our defenses as a pathogen or living in the soil and competing for a new niche near a tree, a bacterium’s genome can mutate in many directions. Where should it begin? Escherichia coli, a common bacterium in our gut that is widely used in research, has a genome of 4.7 million Gs, Cs, As, and Ts, strung in a row. When it faces the challenge of a changed environment, such as when it infects a new person, E. coli has 14.1 million ways to change one single letter, to replace any one of its 4.7 million As, Gs, Cs, and Ts with one of the other three. If each single choice of a letter change were a word in a book about the same size as this one, the book would have over 140 volumes. And what if 20 specific changes are needed for fitness in a new environment? Trying all combinations of 20 changes to get the helpful 20 would require a mass of E. coli that would weigh more than 6,000,000,000,000, 000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,00 0,000,000,000,000,000,000,000,000 Earths. Because each letter in its genome can either stay the same or change to any one of the other three letters, 4.7 million Gs, Cs, As, and Ts strung in a
71
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
72
DARWIN IN THE GENOME
row can create 1 followed by nearly 3 million zeros different possible genomes, not even counting times when letters are inserted or deleted. This list of options is so long that it is hard to describe. And what about our genetic options? Our human genome has more than 3 billion base pairs. For a given E. coli to explore placing each of the three other bases at each position in its genome would require 102.8 million (1 followed by 2.8 million zeros) genomes. E. coli can make copies of its genome very quickly, sometimes even three times an hour, one into two, then each into another two, then each into another two—eight new E. coli from each one every hour. Within 18 hours, a single well-fed E. coli could divide to be your size and shake your hand. From Friday dinner to Sunday brunch, the entire surface of the Earth would be knee-deep in the progeny of this single well-fed E. coli. If this E. coli had started dividing when Columbus got support for his voyage from Queen Isabella, there would have been the requisite 102.8 million E. coli about the time the Bill of Rights was added to the U.S. Constitution 300 years later. But first, the bacteria would have run out of food. In fact, the problem is that if E. coli could divide, in health and without competition, from breakfast on Monday to dawn on Thursday, there would be more E. coli than all of the elementary particles in the universe.2 So E. coli cannot randomly try every possible change and then wait for selection to capture the very best ones. If E. coli got a little fitter each time it made one change that was on the route to a fitter genome, it might get there; but if a single change in the right direction didn’t provide any particular advantage, this E. coli’s descendants might never reach the fitter bacterial genome. It is as if the bacterium faces a huge landscape full of alternative new genomes. Which ones should it approach? Genomes that evolve efficient biochemical systems to navigate through the space of possible future genomes would have an advantage. They and their descendants would discover, more quickly, the direction to take, a way to adapt to an environmental challenge. These bacteria would race ahead of those that are wandering aimlessly around the genomic landscape, those for which evolutionary innovation awaits the results of purely random mutation. The bacteria’s repertoire of options is not limited to a high mutation rate, or even to a high mutation rate focused just in certain genes. Sometimes, even for a bacterium, it is more efficient simply to ask a neighbor for advice. Why reinvent the bacterial version of the wheel? Bacteria can capture genetic information, the gift of other genomes’ experience, by getting DNA from the outside and pasting this DNA into their own genomes.
EVERYONE HAS SOMETHING TO TEACH US
73
Our first hint that bacteria can learn from one another came from a mystery involving bacteria responsible for pneumonia. Some pneumococci bacteria were deadly, and some were harmless. When the deadly pneumococci were killed by heating them in the laboratory, not a single one survived, but somehow they were not fully dead. When harmless pneumococci were mixed with material from the heat-killed deadly bacteria that had been shaken in soapy liquid, the harmless bacteria became deadly. In addition, the once-harmless bacteria’s progeny also were deadly, for generation after generation. How could generations of bacteria inherit information from dead bacteria? Oswald Avery wanted to figure this out. At the Rockefeller Institute in New York City, Avery and his colleagues Colin McLeod and Maclyn McCarty worked to purify, from the many things in a soup of heat-killed, soap-shaken bacteria, the “transforming principle.” They wanted to find what it was in that soup of pureed dead bacteria that could transform harmless pneumococci and their descendants into killers. Year after year the Avery team grew the deadly bacteria, killed them, used a converted cream separator to separate the killed deadly bacteria from the fluid they grew in, and dissolved the killed deadly bacteria in soap. The soapy dead bacteria were then put through something like a carnival ride, a centrifuge, which spun them around at high speed, so that the heavier things flew to the outside end of the test tube. The transforming principle—whatever it was—remained dissolved in the liquid. The researchers then poured the liquid that held the transforming principle into another test tube and added ethanol. The ethanol caused the transforming principle, to crash out of solution into a pellet along with some, but not all, of the other molecules that were in the mixture. The Avery team poured out and discarded the things that stayed in solution above this mysterious pellet; then they added salt water. This took the transforming activity back out of the pellet and into solution again. On and on they went, separating the molecular contents and testing pellet and soup to find where the transforming activity had gone. It was as if they were working with a bowl of alphabet soup with letters too tiny to be seen, trying to isolate the invisible letter P. After years of work, they had it.3 What Avery and his colleagues found was completely unexpected. To everyone’s surprise, they found that the transforming principle that could come out of killed deadly bacteria and transform harmless bacteria into killers was DNA.4 This was in 1944, and back then everyone assumed that DNA was a simple molecule, limited to some kind of scaffolding role in the chromosomes.
74
DARWIN IN THE GENOME
After all, DNA is made up of only four “letters”: A, T, G, and C. Even simpler than that, the number of As always equals the number of Ts, and the number of Gs always equals the number of Cs.5 So DNA was made of only two kinds of things, even amounts of A and T and of G and C, and yet now it seemed that it was acting as if it were the stuff that genes are made of—encoding information that is passed down from generation to generation! How could this simple molecule carry so much information? As Avery himself wrote to his brother Roy, “Who could have guessed it”?6 Many people still doubted that DNA could do this, even after Maclyn McCarty and Jacqueline Jonkowske showed that an enzyme that destroyed DNA also destroyed this transforming principle. This was treated as so incredible that there was no Nobel prize awarded to Avery and his team, even though these careful, thoughtful workers had made one of the major discoveries in human history: that the genetic material is made of DNA. Avery and his colleagues were, of course, right. In fact, these very experiments with pneumococci were the first clear demonstration that genes are made of DNA. DNA, a chemical, a translucent, shimmering substance that can be spooled up out of a test tube onto a glass rod, could, by itself, transfer information from one bacterium to another, and this information could somehow be copied and transmitted to descendants. The focus on DNA that followed led within a decade to the publication of Watson and Crick’s famous model, based upon Chargaff’s rules (A = T and G = C) and Rosalind Franklin’s work with her student Raymond Gosling showing that DNA is a double helix with the letters next to each other. (Unknown to Franklin, Watson had peeked at, and essentially stolen, her data before she, a very deliberate worker, had a chance to present her interpretation of what it meant.7) The Avery team’s discovery that DNA from deadly bacteria could transform harmless bacteria and their descendants into killers for generation after generation was a double surprise. First, there was the surprise that such a “simple” molecule as DNA could be the stuff of genes. Second, there was the inescapable conclusion that genes can pass not just from “parent” bacteria to “progeny” bacteria (vertically in a family tree), but between two bacteria in the same generation (horizontally on the family tree—as if, at least among bacteria, genetic material can pass from brother to brother to cousin to classmate to neighbor). So the very discovery that DNA was the genetic material resulted from the ability of DNA to transfer information, not just between parent and child, but between two different bacteria. Bacteria could “learn” from their environment by swallowing pieces of DNA.
EVERYONE HAS SOMETHING TO TEACH US
75
The ability to learn from its neighbors provides a tremendous leap forward for the bacterium that we left sitting at the edge of the mutation landscape with uncounted options before it, each of unknown value. The advantage that bacteria gain by sampling DNA sequences from other bacteria is demonstrated by their resistance to antibiotics. Indeed, information encoding resistance to nearly all commonly used antibiotics has been found to be strung together in a useful information packet—a sort of “how to” manual for antibiotic resistance that bacteria pass among themselves. A newborn baby had trouble breathing and so was put on a respirator. He had an infection caused by Staphylococcus aureus bacteria, and he was given antibiotics. This baby was in the Netherlands, in the year 1998, but of course this sequence of events is not unusual; it happens to many babies in many places. But what happened next was, we hope, unusual. After being treated with antibiotics, the baby seemed to recover from the infection. But when the baby’s infection came back, the antibiotic amoxicillin no longer worked. Before he was two months old, the baby was fighting a strain of S. aureus that had become resistant to amoxicillin. Fortunately, the baby recovered. The medical staff was determined to figure out where the resistant bacteria had come from. The baby had not traveled. The nurses and doctors and other hospital staff were tested and did not carry these dangerous resistant bacteria. DNA from the baby’s antibiotic-resistant S. aureus was compared to DNA from over 300 samples of amoxicillin-resistant bacteria from all over Europe. It was not any of them, but it did look familiar. In fact, the resistant S. aureus that infected the baby looked very familiar. It was the same S. aureus that had caused the baby’s original infection—the same, that is, except for one small change: A patch of about 40,000 new letters had slipped into the S. aureus DNA.8 Amoxicillin, like other antibiotics in the penicillin family, kills bacteria by interfering with their ability to build their cell walls, an essential structure that surrounds each bacterial cell. The patch of new letters in the newly resistant S. aureus that infected this baby encode a protein that can keep building the bacteria’s cell wall without being bothered by amoxicillin. These 40,000 letters, and thus the information that allowed the bacteria to survive in the baby while swimming in antibiotic, must have come into S. aureus sideways from somewhere. In fact, they came from somewhere inside the baby. Also growing in the baby was a cousin of S. aureus called S. epidermidis. The cousin had this 40,000-letter patch in its DNA, and had passed it over to S. aureus.
76
DARWIN IN THE GENOME
Bacteria help one another resist antibiotics; they also can help one another resist our immune system. Much like the Lyme spirochete’s system of moving patches into its coat protein, bacteria can share information encoding coat patches with one another, providing a greater repertoire than is available in any one strain; this happened in a recent epidemic of meningitis in West Africa, enabling bacteria to make rapid changes in their surface as they spread through the population.9 The S. aureus that infected the baby in the Netherlands already was a pathogen; what was new was that it had become resistant to antibiotics. But bacteria that are not pathogens, that normally live within us while causing no harm, can get new information from other bacteria and become pathogens. This is what happened to the harmless strain of pneumococci in Avery’s laboratory, and it has happened to some E. coli on our farms. E. coli lives with us, harmlessly, in our guts. It is so common, and so harmless, that a strain of this bacterium—coincidentally called K-12—is used in laboratories by inexperienced students for simple experiments as well as by professional researchers, in routine ways, every day. This tame, safe bacterium has been our reliable teacher. Without subjecting us to risk, it has shown us how DNA is copied, how the simple viruses that infect bacteria work, and many other things. But suddenly, one day intense TV anchors were talking about our friend E. coli; it had killed people who ate hamburger at a picnic. Once cattle are infected with this aggressive form of E. coli, then everything from hamburger to unpasteurized milk, to fertilized fruit and vegetables, to ice cubes made from water contaminated by runoff from cattle pastures can make people very sick, and often kill them. In fact, we now know that E. coli sickens about 75,000 people in the United States alone every year. What had happened to our laboratory sidekick? It was as if coming home one day we saw our pet teacup poodle turn into an aggressive pit bull. It is hard to believe that the teacup poodle with the pink bow, and the snarling pit bull are members of the same species, although they both are called dogs. Now that we can read the complete sequence of its genome, we can get a fairly thorough answer to the question of what happened to E. coli, an answer that would have been inconceivable in its detail a few years ago. Two complete E. coli genomes were compared: the genome of the standardissue laboratory companion E. coli, which lives harmlessly in our guts, and that of the killer, a strain of E. coli called O157:H7 that sickened people who ate undercooked ground beef in 1982 in Michigan. The DNA sequence of O157:H7 was lined up, letter for letter, with the DNA sequence of the
EVERYONE HAS SOMETHING TO TEACH US
77
tame E. coli.10 The comparison revealed not only how the two strains are different, but also how they became different. Unlike our genome, which is arranged on two sets of 23 linear chromosomes, the E. coli genome is on one circular chromosome. Most of the way around, it was clear that both circles were E. coli: 4.1 million base pairs of the two E. coli genomes lined up well, like a common backbone, although a piece of the O157:H7 genome was flipped around so that it was backward. This is not to say that the 4.1 million base pairs were identical. We can compare you and your cousin and find that she has green eyes while your eyes are hazel, but your eyes and hers clearly are related structures. Out of E. coli’s 4.7 million letters, there were 75,168 individual changes from one letter to another. On the average, the same gene in the two strains was alike more than 98 percent of the time, although a few genes involved in interaction with the host (e.g., us) were only 34 percent identical. Most of these single-letter changes were between two codons that code for the same amino acid—i.e., between synonyms. These single-letter changes were not responsible for the dramatic change in the behavior of E. coli. The variation introduced into the genome by these single-letter changes pales in comparison to the fact that over a million letters were found that had come into the genome sideways. Scattered around the circles of the two E. coli genomes were hundreds of patches of DNA that were found on only one of the circles, like islands of unrelated DNA in a sea of similar sequences. One of the islands on the deadly bacterium’s genome, containing 106 new genes, came in two identical copies. In fact, there were a total of 1.34 million letter pairs, estimated to contain 1387 genes, that were clustered on islands found only in O157:H7, called “O islands.” There were about half a million letters, with 528 genes, clustered on islands called “K islands” because they were only found in K-12. While the O islands were not found in K-12, they were not completely unfamiliar. In fact, they include many genes known to be, or suspected of being, involved in diseases caused by other pathogenic bacteria. For example, some of the O islands encode proteins that help bacteria stick onto our tissues. Others encode toxins, including a toxin that harms our macrophages (cells that usually engulf and kill bacteria) and a “type III” secretion system that also is found in virulent strains of bacteria like Salmonella and Shigella, which use it to inject things into our cells. These types of genes are nasty, invasive, and manipulative—and they lead to disease. Where did they come from? Some of them seem very similar to genes used by bacteria that get intimate in a friendlier way with the cells of organisms such as plants and animals. These helpful bacteria stick
78
DARWIN IN THE GENOME
to plant roots; indeed, they invade root cells and stimulate their growth. They also take nitrogen out of the atmosphere and convert it to a chemical form that can be used by plants. All plants and animals depend for survival either directly or indirectly on these bacteria; the nitrogen that is converted, or “fixed,” with the help of these bacteria is essential to them, and it turns up in their, and subsequently in our, proteins and DNA.11 Examining E. coli gives us a view of evolution that differs dramatically from the model of random letter-by-letter mutation followed by selection. Bacteria are not restricted to that unstructured approach. How do bacteria learn from their neighbors without the cream separator, centrifuge, and other tools used in the Avery lab? Bacteria have evolved, and no doubt have acquired, mechanisms that enable them to sample and adapt information that is available in their environment and, in a two-way exchange, to share information with neighboring bacteria by sending pieces of DNA outside. DNA uptake is not an accident. Originally, perhaps, a bacterium, exploring for food, with an ancient, molecular curiosity, had taken up a dead bacteria’s DNA and discovered what it could do. Modern bacteria have multiple mechanisms that they can use to acquire DNA. Among the gene-swapping tools used by bacteria are integrons, a genetic framework that carries genes and groups of genes in and out of bacterial genomes.12 Integrons make it easy for a bacterium to pick up not just fragments of DNA that happen to be lying around, but entire genes, and indeed to pick up and broadcast whole sets of genes. New genes can pop in and out of integrons like cassettes. Unfortunately, these cassettes do not play beautiful music. Instead, they spread antibiotic resistance, pathogenicity, and other properties. Integrons also encode their own enzyme for inserting genes into and removing them from other DNA, and information that allows their genes to be turned on and copied into RNA messages to make proteins. Integrons have tremendous significance for the emergence of new pathogens. Some integrons carry information into willing bacteria that helps the bacteria to survive in the face of nearly all known antibiotics. In fact, learning from their neighbors is so important to bacteria that they have many mechanisms available to them, not just one. Bacteria can swallow naked DNA, they can take up “packaged” DNA, and they can have sex.13 Bacteria can learn from other species. Indeed, rather than prospering by shoving out other strains in a competition for survival, bacteria may benefit when other strains live near them, inside us, for this adds to the diversity of DNA available to them. Bacteria can swallow naked DNA that happens to be in the soup around them, say from neighborhood bacteria that spilled their contents when they died.
EVERYONE HAS SOMETHING TO TEACH US
79
The bacterium Haemophilus influenzae preferentially takes up DNA from other members of its own species. (Some strains of H. influenzae live harmlessly within us, but others can cause life-threatening infections, from respiratory distress to meningitis, especially in children.) The naked DNA is taken into a protected compartment at the bacteria’s surface. Uptake of DNA into this compartment is facilitated by a “password.” DNA passes easily between bacteria of this species because the DNA knows the password. In this case, the password is a nine-letter-pair sequence: AAGTGCGGT TTCACGCCA If you looked for the password in the first strain of H. influenzae that was sequenced, called Rd, you could see that it is likely that any piece of H. influenzae Rd DNA that is floating by will carry the password, as there are 1465 copies of the password in each copy of the bacterium’s single 1.8-million-letter-pair chromosome.14 Following its uptake, the DNA slowly moves out of the protective compartment into the cell itself. When it enters, one strand of the helix is completely chopped up, and the other strand is partially chopped up. This may seem like an inauspicious welcome, but Nobel laureate Werner Arber suggests15 that perhaps it is safer for the cell to chop up the DNA into bitesized pieces than to allow a whole genome to enter and take over. If a piece of DNA can find a good patch of matches on the host bacterium’s DNA (matching by the A-T G-C pairing rule), the entering strand will push aside one of the host’s DNA strands, recombining and sharing the information that it has. Bite-sized pieces binding similar sequences are most likely to vary a preexisting gene, making many changes at once. Since these changes come from a version of the gene that is functioning in the bacteria that sent the DNA, they contain valuable information that can guide the host cell through the mutation landscape. Something like this happens to our own genome when we form sperm or eggs, as is discussed further on in the book. Passwords make information sharing more likely within one species of bacteria, but bacteria share DNA with other species, too. New genes can be made this way. In one bacterial genome, the gene encoding an enzyme that digests a carbohydrate was pieced together out of DNA from two different species;16 the part that sticks to the carbohydrate came from one species, and the part that cuts carbohydrates from another. So, with help from their neighbors, bacteria can learn to digest new food sources. While H. influenzae often swallows DNA, other bacteria more often exchange genes through extra pieces of DNA called “plasmids,” using sex,
80
DARWIN IN THE GENOME
or via bacterial viruses. Not likely to find its way onto the screens of movie theaters in red-light districts, bacterial sex usually occurs in settings that are outwardly more innocuous, such as an idle food processor.17A bacterium can build structures that facilitate the transfer of a copy of its DNA into an attached bacterium. Bacterial sex may increase under stress, which turns on genes that create the environment that allows the DNA transferring system to be built and to function. Another way in which bacteria can get new genes is with the help of viruses.18 One type of bacteria, which can grow in both insects and plants and which damages citrus trees, got at least 83 of its approximately 2900 genes, making up 7 percent of its genome, from viruses; many of these genes are responsible for the bacteria’s virulence.19 Another type of bacteria can become virulent and cause cholera when it gains a cluster of genes by horizontal gene transfer and also from a bacterial virus.20 Viruses are pieces of genetic material wrapped in protein. There are viruses that invade and kill bacteria—(known as bacterial viruses), just as there are viruses that specialize in harming us. But many viruses do no immediate harm when they put their package of DNA into a bacterium, even when they insert their DNA into the bacteria’s chromosome. They may do nothing until the bacterium is stressed, such as with DNAdamaging ultraviolet light, when these viruses emerge again, spreading the genes they carry among other stressed bacteria. Often, the edges of DNA islands have patches of DNA sequences that look like bacterial viruses and encode machinery that can move DNA horizontally from one bacterium to another. These patches also have a different “color,” or dialect—(the relative use of A-T and G-C letter pairs), which indicates that they probably came in sideways from other species. Toxin genes are hooked up to the bacteria virus system in such a way that when stress—such as from certain antibiotics—induces the viruses to leave, the toxins may be expressed too. This is why it sometimes is dangerous to treat a bacterial infection with antibiotics; when the toxin is turned on, it may kill us. On the other hand, not treating an infection also can kill us. Salmonella makes major use of viruses, which provide it with “a transferable repertoire of pathogenic determinants.”21 Salmonella’s ability to infect different hosts depends on the viruses that have infected it. One tiny virus protects Salmonella from attack by an oxidative burst by phagocytic cells. If we remove this gene from the virus, the phagocytes22 can kill these bacteria. The different symptoms of Salmonella infection also depend on the viruses that have infected the bacterium. Macrophages (“big eaters”) defend us against infection. They wrap themselves around bacteria, forming a phagosome (“eaten body”) inside of
EVERYONE HAS SOMETHING TO TEACH US
81
them, where the bacteria are isolated. Then they move a vessel full of digestive enzymes and hydrogen peroxide-generating systems into the phagosome to kill and digest the bacteria. So, the last place a bacterium would want to live, it would seem, would be in a phagosome, where it risks a bath of peroxide and other deadly concoctions. Yet some bacteria seem to live very happily in phagosomes.23 Of course, natural selection would favor the bacteria that came up with a mechanism to block the macrophage’s poison cocktail. This would let them live within a macrophage, hidden from the immune system’s surveillance. Bacteria can pick up, from other bacteria, instructions that enable them to live inside a macrophage; these instructions come packaged on pathogenicity islands. A pathogenicity island picked up by some Salmonella typhimurium prevents the peroxide-generating systems from ever getting into the phagosome where the bacteria were taken after they were swallowed by the macrophage. These bacteria have a type III protein secretion system, which lets them inject some of their own proteins into the macrophage after they are swallowed, interfering with the biochemical machinery that would otherwise kill them. Avery and his colleagues’ discovery that we can change the genes of bacteria by feeding them DNA eventually transformed genetics research. This discovery was the first seed that, within four decades, became genetic engineering. If we need large amounts of human insulin to treat diabetics, we can put the gene for human insulin into bacteria and let the bacteria divide and copy the genes; the bacteria and their progeny will make human insulin for us. Like living, growing, dividing factories, the bacteria construct more and more of the protein that interests us, mixed in with their own. We separate the factory product that we want, in this case insulin, from the others. If we want to learn how to block the growth of a deadly virus without risking our lives by growing the virus itself, we can put a gene that is essential for the growth of the virus into bacteria under careful constraints, get the bacteria to churn out large amounts of viral protein, and use the manufactured protein to study how to block the biochemical action of that isolated protein—safely, away from the deadly virus—in order to kill the virus.24 We can even chop up the human genome, put it into bacteria, and have the bacteria grow extra copies of it for us to study; indeed, that is how we were able, bit by bit, to determine our genome’s sequence of 3 billion letters. So Avery and his colleagues’ experiments with DNA and pneumonia were the seeds of a revolution in research that led, later in the century, to the ability to manipulate, study, and sequence genomes. But, surprisingly, the implications of these experiments for evolution were not really incor-
82
DARWIN IN THE GENOME
porated into evolutionary theory. Where, after all, did we get the biochemical tools that we use in laboratories for genetic engineering? From bacteria. From ancient bacteria that grow in hot springs at Yellowstone; from bacteria that grow in our guts. From bacteria, companies that supply our laboratories have put together, and sell, whole catalogues of molecular scissors that cut DNA in precise ways, molecular paste that links DNA in precise ways, and molecular machines, polymerases, that copy genetic material over and over again. We are not the first genetic engineers.25 That title belongs to the tiny microbes. So, what were the microbes doing with all these tiny tools? Few people asked.26 In spite of the Avery experiments, evolution continued to be viewed in purely vertical, parent/progeny terms. Now that we can look at complete genomes, as we read along a bacterial chromosome, suddenly we find a whole block of DNA that looks different, out of place; it doesn’t fit—there is a cobalt patch on a turquoise string. The relative amounts of A-T and G-C along the helix seem to change suddenly, as if it’s talking a different dialect. A block of genes seems to be very closely related in sequence to the same block of genes in a very different bacterium, encoded by a cobalt string. These genes seem to have come into the more turquoise genome sideways. In spite of the way DNA’s role was discovered, it wasn’t until we had sequenced entire genomes that the prominent role of horizontal gene transfer was incorporated, deeply, into our understanding of evolution. The genomes are shouting at us; we can’t ignore them now. Patches of DNA— genes—have come from somewhere else. Perhaps we didn’t visualize this bacterial networking before because we can’t truly see what we don’t understand. We can’t get the right answer if we can’t imagine how to ask the right question. Perhaps only now that we are networked and internetworked can we see that bacterial genomes are networked too. Just as we post and download recipes from the Internet, bacteria broadcast and receive information on how to destroy or otherwise get rid of an antibiotic, or of all known antibiotics. They have been on the planet a very long time, and they are masters at what they do. Whole organisms have come together in many creative ways to build a new form of life that can do more than any one of these organisms could do alone. It even is possible to evolve, in the laboratory, an amoeba that is completely dependent upon being infected by a specific bacteria in order to survive; after an initial devastating infection that killed many amoeba, the bacteria and the amoeba learned to work together; they became a team.27
EVERYONE HAS SOMETHING TO TEACH US
83
And we, too, live in partnership with other genomes. I want to reflect here on a group of bacteria, cousins to each other, that chose to head down different paths in their relationship with us. Both live within us. One cousin gave up its ability to be free in exchange for a steady supply of amino acids and other foods from us. In order to survive, it must live within our cells. It is spread from person to person by lice; this cousin, typhus, helped tremendously by war, killed tens of millions of people in the twentieth century, including Anneliesse Frank.28 The other cousin chose to team up with an ancient cell; it came to stay. It too lives within us, not as an enemy, but as a partner; in fact, we cannot live without it. The “good” cousin became our cells’ energy factory, our mitochondria.29 We rely on the energy it gives us every time we move an arm or think a thought, and every time our heart beats. For this teamwork, eukaryotes take the good cousin wherever they go on the planet. The good cousin thrives in the sunlight; it is not limited to struggles in disease, war, and famine. Cooperation provides options for moving beyond barriers; it is much more efficient than random change. An infrastructure that facilitates cooperation lets two genomes combine the benefit of what they have learned in their separate experiences. It predates the successful adaptation of cooperation in human society; each of us separately does not have to figure out how to build an airplane when we want to fly. Bacteria do not each have to figure out, separately, how to combat a new antibiotic. It is evolution’s clear message that the fittest is likely to be the most cooperative.
This page intentionally left blank.
8 The Genome Sends an SOS Some people in the crowd wake up . . . they emerge according to much broader laws. They carry strange customs with them. . . . The future speaks ruthlessly through them. —Rainer Maria Rilke
I
n 1967, Evelyn Witkin, a geneticist then working in Brooklyn, New York, wrote two papers that at the time seemed to be on different subjects.1 Both were based on what happened when she gave Escherichia coli a bit of a sunburn. The first paper examined how ultraviolet (UV) light affected the behavior of the bacteria, and of a virus that had inserted itself into E. coli’s genome and was quietly passing from generation to generation along with the bacteria’s DNA. When the bacteria were sunburned, they began to grow in long threads rather than separately, and the dormant virus was “induced”: It started to reproduce and spread through the colony of bacteria. Witkin reasoned that damage from the ultraviolet light generated a signal within the bacteria that had a variety of effects, including changing their growth behavior and awakening the dormant virus. Witkin’s seemingly unrelated work in the second paper was about mutations. Usually, ultraviolet light causes many mutations. However, when she shone UV light on bacteria that already had a mutation in a place in their DNA called LexA, she didn’t find these UV-caused mutations. Previously, as Witkin put it, “the prevailing notion . . . was that mutations were instantaneous events . . . the mutagen went ‘zap!’ and that was that.”2 In contrast, Witkin’s data suggested that to cause these mutations, UV light actually
85
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
86
DARWIN IN THE GENOME
needs the help of the bacteria themselves, in particular of LexA. Witkin suggested that LexA’s role was to provide a way for bacteria to copy their UV-damaged DNA. While the bacteria without LexA couldn’t copy their damaged DNA and died, the bacteria with LexA could copy their DNA, but this copying was error-prone and so caused mutations. The survivors Witkin found that did not have UV-induced mutations were the progeny of bacteria that either had managed to repair their DNA without error before attempting to copy it or had escaped UV damage in the first place. So, here is the situation. Suppose you are the captain of a ship, and you have on board a well-trained engineer who can, with a box of precise tools, perform routine maintenance on the ship, patching occasional damage so well that no one who inspected the ship could see that it had been patched. Also on board is a group of very enthusiastic and inventive do-it-yourselfers who can fix just about anything, but who leave rough edges and loose ends—in general, a bit of a mess. Whom would you rely on? Well, there’s one more piece of important information. The exacting engineer has a personality quirk: If the damage is so severe that, working with these precision tools, a perfect repair cannot be made, this engineer will refuse to touch the spot. Now suppose that your ship bangs into a rock, and a three-foot gash is opened just at the water line. The engineer backs away in horror. Will the ship go down, the damage unrepaired? In rushes the team of do-it-yourselfers. A metal plate, some rags, a carton of Krazy Glue that someone found in cargo—whatever it takes, the ocean is kept out and the ship is saved. It may not be the same sleek ship it was before it hit the rock, but it will be able to continue its journey. So, too, for DNA. The precise DNA copying machinery demands an exactly matched pair of letters, A and T or G and C, before it can go on. If this machinery puts in the wrong letter, it cuts this mismatched letter right out. It cannot, except ever so rarely and only accidentally, extend misshapen DNA. We depend upon this personality quirk, for this fussiness is the basis of our ability to copy our 3-billion-letter genome accurately, time and time again. This machinery is intrinsically unable to be sloppy. Missing or unrecognizably damaged letters must be replaced before DNA is copied, or the copying machinery may stall or fall off. Will the cell die? Perhaps many cells did die under such circumstances; they have no descendants among us today. Miroslav Radman, the son of a fisherman and currently a molecular geneticist working in Paris, recognized the nature of this emergency.3 Radman recognized that cells would evolve a response that would save their life—that would keep the ship from going down— when DNA copying stalls, and he named this the SOS response.
THE GENOME SENDS AN SOS
87
So it turns out that Evelyn Witkin’s two papers were very much related to each other. In response to DNA damage, the sunburned bacteria turned on the SOS response. This had systemic effects on the bacteria—on the way they grew, on the behavior of the viruses that were asleep in their genome, and also on their ability to copy damaged DNA. A central player in the SOS response is a protein called RecA. When the copying machinery falls off, RecA sticks to the separated single ribbon of the DNA helix. What RecA does next determines how the damage is fixed. If a photocopy came out smudged in one spot, you could look back at the original to read the correct letter. Or, if a spot of water smudged the original, you could find the information on the unsmudged photocopy. You could use the same approach if you were a repair system sent in to fix damage to DNA, and this is what RecA does. Sometimes RecA can get the damage repaired by recombination with another DNA molecule, perhaps the very DNA molecule that contains the damaged strand’s recent partner, before copying. But if RecA, holding onto the piece of damaged DNA, cannot quickly find a good match to repair the damaged region, it will send out the alarm:4 The copying machinery fell off, the cell could die—Save Our Ship. As the first signal in the alarm, RecA shoves another protein, causing it to cut itself;5 this other protein turns out to be none other than LexA. We now know that LexA is a repressor protein, something that sticks to genes and keeps them turned off. It is as if LexA were a climber holding onto a cliff and RecA were banging on its fingers. LexA loses its grip and falls off, allowing about 20 to 30 genes that it had repressed to turn on. Among the genes that are turned on when LexA’s repressive grip is released are genes that encode enzymes that can copy damaged DNA. Because DNA is made up of a string of letters, it, like nylon, is a polymer of repeating units. The ending -ase refers to its role as an enzyme. Thus, an enzyme that makes a polymer is a polymerase, and an enzyme that makes DNA is a DNA polymerase. These polymerases can be activated and directed to the damage by RecA.6 In fact, there isn’t just one enthusiastic do-it-yourselfer polymerase. In E. coli, we know of five, each of which has slightly different skills.7 One of these can copy DNA that is so badly damaged that its letters are unrecognizable. It can even copy DNA at spots where a letter is missing—it makes things up, it plays the odds. One enzyme simply adds a C across from a place where the DNA is missing a letter. If the DNA is still a bit misshapen after only one letter is added, this enzyme can add another C. This gets us past the damage8—it saves the ship—but it does not make an accurate
88
DARWIN IN THE GENOME
copy. In other words, it causes mutations. Working together, these polymerases can fill in the hole where the DNA was damaged and extend the chain past the damaged spot so that the careful polymerase can step back in to complete its job. This is what happens in bacteria like E. coli, but what happens in us? Much of the important biochemical machinery in our cells, including DNA itself, emerged in an ancient organism that was an ancestor both of us and of the bacteria alive today. Therefore, it is fairly standard in biology to develop hypotheses about what might be happening in our own cells based on the biochemistry of bacteria. These hypotheses need to be tested, of course, often first in yeast or in little animals like fruit flies and worms. But now that we have whole genome sequences, we can actually peek inside to see whether there are enzymes encoded in our own genome that look like the bacteria’s unusual polymerases. There are.9 When one of our own versions of these SOS polymerases copies undamaged DNA in the laboratory, it makes a mistake about once in every 200 letters it copies.10 That would be more than 10 million mistakes every time it copies our several-billion-letter genome. Clearly, copying undamaged DNA is not this enzyme’s day job. Our DNA is not kept in an impenetrable protected vault. It moves through life with us every day. If you are reading this book on the beach, I am sorry to bring this up, but ultraviolet light tends to tie together two Ts that are next to each other on the same side of the helix—a T dimer; this jams up the DNA copying machinery. UV light from the sun struck the first DNA on the surface of the Earth. We have been living with UV light for so long that we are not passive in the face of this T tying. We have enzymes that spot and cut out T dimers. If the dimer has not been cut out by the time the DNA is copied, the standard polymerase cannot copy it. Fortunately, we have another enzyme that can put two As in the new strand of DNA across from the funny misshapen piece of DNA with the two Ts tied together. This is more than a good guess. A T dimer comes from two Ts, which means that an accurate copy of the DNA as it was before the damage should have two As on the new partner strand. The enzyme that can copy T dimers does this reliably, much as the standard enzyme recognizes the normal Ts one at a time. However, if this enzyme is let loose on normal DNA, it makes a lot of mistakes, getting about one letter wrong for every 18 to 380 letters copied.11 And it can’t proofread. There are other ways of getting past a T dimer, but they cause mutations. People who are missing the enzyme that can repair T dimers are likely to get skin cancer from UV radiation,12 for however their DNA-copying
THE GENOME SENDS AN SOS
89
systems may deal with the T dimer, they do not reliably insert two As across from the damage. Even when the T-dimer-fixing enzymes are intact, different sequences of DNA letters near the dimer can affect how well the enzyme can get in and fix it. On the other hand, as Evelyn Witkin observed in her sunburned bacteria, some mutations actually decrease the amount of DNA damage that we see among surviving progeny. If a particular polymerase that copies damaged DNA is inactivated in human cells in the laboratory, little or no DNA damage will be induced by UV light, but these cells will be slightly more likely to die. In other words, knocking out this polymerase may protect against cancer because the cells will die rather than mutate.13 Even aside from UV light, mutation is a constant threat. Genomes have a natural tendency to fall apart. As a whole, DNA is stable, but, truth be told, it is not as stable as your living room chair. It constantly is dropping things and constantly being repaired by ever-vigilant proteins. DNA is a high-maintenance material. Every day in every cell in each of our human bodies, 5000 As and Gs fall off the DNA backbone just from bouncing around at the warmth of body temperature.14 Fortunately, these missing letters can be replaced accurately, using the information in the complementary strand in the helix. Then there is the problem of restless letters. Like a cat stretching, every now and then the bonds between the atoms that make up the letter C can shift around a bit, so that just for a split second the C looks like a T. C stretches occur so often that in a genome with nearly a billion Cs, as many as 100,000 of them could happen to be stretching into Ts as the copying machinery rushes by. This stretching is so quick that as soon as the copying “mistake” is made, the C is back to itself. Proofreading then cuts out the A that had been inserted to pair with the momentary T, which now is inappropriately across from a C, and puts in C’s partner G, and the polymerase moves on before the C stretches again. But stretching occasionally to look like a T isn’t the only way that the letter C can change into a letter that pairs with A rather than G; C also tends to lose a piece entirely, turning into the letter U. There is an enzyme that is on the lookout for Us in DNA, but if the broken C is not replaced before it is copied, the new DNA strand will get an A at that spot, not a G. Then, if the A is copied, it will bring in its partner, a T. This removes all trace of the original C, as if “lace” had been changed to “late.” Is this mutation rare? Every day, in each of our cells, about 100 of the Cs lose this piece and become Us. Cs have been losing pieces and turning into Us since the very first DNA molecules appeared. A repair system has evolved that spots Us and cuts them out of DNA, also chewing off a few letters on either side to
90
DARWIN IN THE GENOME
make room for the DNA copying machinery. The copying machinery sees the C’s original G partner across the double helix and fills in the gap, putting the C back where it belongs. DNA is a molecule that is touched by all of the effects of its environment. C is not the only letter that gets damaged from day to day. Groups of atoms can fall off of other letters too, and extra atoms can get stuck onto letters. There are about 20 enzymes that recognize different common types of damage to DNA and fix it. Between letters that fall off and enzymes that cut off damaged letters, it is estimated that in each of our cells every day, about 9000 places in the DNA backbone are missing letters.15 Each cell is kept busy patching its genomic dike. The number of places in the DNA that are missing letters at any moment varies from one cell to another by over an order of magnitude. Since whether these mutations are fixed depends upon the activity of enzymes, how much they are fixed can be different at different places in a genome. Whether or not this is the explanation, it is true that in our genome, some regions are richer in As and Ts, and some in Gs and Cs. Mutations can happen even where letters haven’t fallen off, if the exacting polymerase errs as it rushes past. A mismatch between the template strand and the new strand most likely will be caught by the polymerase’s proofreading activity as DNA is being copied. The proofreader assumes that the information in the template is correct. If there is a G-T mismatch and the G is on the template strand, the T will be cut off and replaced with G’s proper partner, a C. Even if proofreading misses an error, mutations can be avoided by the action of a repair system that follows the polymerase along the double helix. But how can the repair system identify which is the template strand? If it notes that a G is paired with a T, or that a few letters are unpaired and looped out, which is right? Which strand was the original? Should the G-T pair be fixed to a G-C pair or to an A-T pair? Should unpaired letters be copied or cut out? Which strand is wrong? One way to keep track of which strand came from the original helix is to mark both strands of the original DNA helix before copying them, like painting a patch of purple on every apricot step of the magic staircase in the plaza. When the strands of the “marked” DNA double helix separate and are copied, one strand of each new double helix will have come from the original helix; that strand will be the one with the purple mark. While E. coli doesn’t have a can of purple paint and a tiny paintbrush handy, it does have a way to tag the original double helix. It marks the DNA by attaching a specific chemical label to certain letters. This molecular equivalent of purple paint, a carbon atom bound to three hydrogen
THE GENOME SENDS AN SOS
91
atoms, called a methyl group, guides this bacterial repair machinery: This strand is the original.16 In addition to a “purple paintbrush” enzyme, many proteins are needed to spot and repair mismatches in DNA. One of these proteins binds to a place on the double helix with a funny shape, whether that shape is a mismatched letter or a place where extra letters were stuck in or letters were lost, so that one strand is longer than the other and a loop sticks out. Another protein finds the marks on the original template strand. Then yet another protein notes the proteins that mark the mismatch and the template strand, and cuts the new strand on the left side of a specific sequence of letters nearby (in E. coli, on the left of a nearby GATC). Yet another mismatch repair protein starts from the cut and chews in past the mutation, leaving behind a stretch where the double helix is missing one of its strands. Finally, a polymerase fills in the missing letters, restoring the double helix using information from the original template strand. When Meningococcus is missing the “purple paintbrush” enzyme it uses to guide mismatch repair, its coat changes more quickly.17 If the mismatch repair system isn’t working, one strand may end up with an extra copy of some letters. If the polymerase fell off and got back on in the wrong place, there might even be extra copies of one or more entire genes. When a mismatched double helix is copied, two different double helices may be sent on to the next generation. Bacteria that are missing mismatch repair are called mutators. Evelyn Witkin found mutators when mutations knocked out the brakes on the SOS system, leaving its unusual polymerases stuck “on.” Is it dangerous to be a mutator? Mutation can be seen as a generally harmful random attack on a welladjusted genome. After all, any species that is alive today has been selected to handle life in the range of habitats it typically encounters. But when an organism gets out of its comfort zone and is stressed to its limit, its progeny need to evolve in order to survive. If a bacterium is struggling, extra copies of some genes that it already has may help it survive until another mutation solves the problem.18 It has to mutate. In fact, compared to pampered laboratory strains of E. coli, samples of “wild” E. coli, which have to handle repeated stresses, whether within people or elsewhere out in the environment, contain bacteria that tend to have a higher mutation rate.19 Radman suggests that turning on SOS mutation is the genetic equivalent of joining with your extended family in a lottery pool, increasing the chances that one of the many progeny could have the lottery ticket to the future.20 While mutators are at a disadvantage under comfortable circumstances, they can have more surviving progeny under challenging circumstances.21
92
DARWIN IN THE GENOME
We have more genes than we use at any one time, as a back-up, to handle the range of fluctuations in the environment in which we can live. The ability to sense and to respond to stress has evolved, enabling cells to adjust, for example, to changes in available nutrients. Bacteria that are fed the sugar lactose will draw on information in their DNA that encodes proteins that absorb and digest this sugar—unless the more readily usable sugar glucose is abundant. If glucose is abundant, why waste energy making the extra proteins needed for digesting lactose? But for every organism, there will be a point beyond which it cannot survive unless its genome mutates. As we go up above sea level, our body adapts to the decreasing oxygen in the air by making more red blood cells. But we cannot go above a certain altitude, which varies for different individuals, without a pressurized cabin (or a spacesuit). So, too, bacteria may find themselves in a challenging environment. Perhaps a droplet containing bacteria from a mouse landed in your nose. If a mouse bacterium can mutate to expand its progeny’s territory to include the human species, those progeny could live in all those places that are low on mice but full of people; maybe someday they could even fly across the world in one of those pressurized cabins. Biological systems are constantly regulating and adjusting themselves in response to changing circumstances. Because of the bacteria’s own role in mutation, from the care of their polymerases to the action of mismatch repair, selective pressure acts on mutations. Increased activity of unusual polymerases and decreased activity of mismatch repair are examples of changes that can increase the chance that there will be many additional mutations. Changes in the activity of genes also can lead to greater acceptance of DNA that comes in from the outside. A successful genome evolves an optimal balance between the competing needs to be stable and to change, between genetic fidelity and genetic exploration. But what is optimal is not the same for all genes in all circumstances. As a biochemist, looking at the sequence of an organism’s genome, I could wire things up to increase SOS polymerases and decrease mismatch repair under stress. I could do this, but to what extent has natural selection done this? Right now people disagree strongly about whether or not certain experiments show that it has, but eventually this question will be answered by careful analysis of the wiring of genomes. While it may seem strange to consider the idea that mutation could be adjustable, at first glance the ability to adjust the mutation rate under stress does sound useful. An efficient pathogen might increase its genetic variation during the first hours when it infects a new host and is trying to
THE GENOME SENDS AN SOS
93
get a foothold.22 But on the other hand, when the mutation rate is increased, it can damage a lot of genes that the bacteria need if they are to remain viable. It would be much better, from a pathogen’s point of view— (or worse, from ours), if mutations became more likely in just those spots in the pathogen genome where variation is least likely to harm the pathogen and most likely to help it adapt to us. We know that there are such spots, such as slippery DNA. Indeed, distinct types of mutation can become relatively more or less likely when different proteins are present in the cell. Changes in the type and extent of mutation do not have to be randomly distributed throughout the genome. When mismatch repair is damaged, some mutations tend to increase more than others, and which mutations increase more can differ depending upon which mismatch repair proteins are inactive.23 Without mismatch repair, changes in the length of slippery DNA become more frequent. In yeast with damaged mismatch repair, changes in the length of a particular TTT became one-third of all mutations detected in one gene, whereas normally only 1 in 20 mutations detected were slips at that TTT.24 So, too, when the activity of a polymerase that doesn’t proofread was increased in mouse cells, mutations in undamaged DNA increased tenfold.25 One-third of these mutations were slips, turning the “tuning knobs” faster. Not all slips are strategic. There was nothing obviously strategic about the TTT slips. While there are many places in a genome made up of slippery and other more mutable DNA sequences,26 the E. coli genes that encode proteins that respond to stress appear to have more than their share of distinct classes of more mutable DNA.27 These more mutable genes include those that repair mismatches and those that repair damaged letters. The DNA encoding mismatch repair gene seems to experience a lot of cutting and pasting,28 for its sequence varies between bacteria in a very patchy way. Mutations in DNA within the genes needed for mismatch repair result in an increase in uncorrected slips and other mutations elsewhere in the genome. In challenging circumstances, selection may favor bacteria that are less efficient at mismatch repair. 29 If mutations overcome the problem and relieve the stress, those among their progeny that regain the ability to repair mismatches are likely to do better. Patchiness may be a footprint of repeated cycles of losing and regaining, through recombination, of active repair genes as the optimal mutation rate rises and falls. This patchiness might result from recombination at closely spaced repeats. Such repeats are more prevalent in these genes than you would expect by chance.30 It appears that distinct types of DNA sequence, with different stabilities, can tend to be selected for where they provide an advantage.
94
DARWIN IN THE GENOME
Among the strategic places where slippery DNA can be found are locations that vary the bacteria’s surface, speeding its exploration of how best to get a foothold in its environment. For a pathogen, that environment would be us. Richard Moxon and his team in England have shown very clearly that, for the bacteria they have examined, a particular kind of slippery site, repeats of four letters, is found in the genome at places where these bacteria tend to focus exploration on genes that contact and manipulate the host. These genes enable a pathogen to figure out how to bind to our tissues, how to get across barriers, how to gain entry to our cells, how to stay ahead of our immune system, and how to overcome other roadblocks to the pathogen’s success. A pathogen that makes all these creative, genome-altering strategic changes while leaving its internal housekeeping genes relatively unchanged and undamaged clearly is at an advantage. Slips in these genes facilitate exploring the ability to bind to us, to poke and prod our defenses. Sometimes they break through. Aside from causing back-and-forth variation of genes that already are inside the bacteria, mutations can affect how much DNA from the outside the bacteria will patch into its own genome. Perhaps some of the neighbors of a bacterium that is stressed by the environment have solved the local problem and so can pass along some helpful genes, such as how to destroy an antibiotic or how to digest a new food source. Because extra pieces of DNA in bacteria, called plasmids, can carry genes between bacteria, genetic changes on plasmids that prove favorable are positioned to spread especially rapidly from bacterium to bacterium.31 One of the SOS polymerases can cause about a 1000-fold increase in mutations in plasmids. The bacterium Enterococcus lives in our guts. Though it is often harmless, it is a major cause of hospital-acquired infections, some of which, including heart inflammation, are very serious. When stressed, Enterococcus can ask its neighbors for help using the molecular equivalent of a note in the form of a string of eight amino acids. A specific plasmid responds to the note by providing the information needed to build the apparatus that enables the plasmid to transfer to the cell that requested its help. Whether the plasmid responds, and also how many extra copies of the plasmid are made in the bacterium, depends upon the number of repeats in the plasmid of the letters TAGTARRR (where each R can be either an A or a G).32 Because repeats can be slippery, information sharing may turn on more quickly when mismatch repair is not working. When bacteria are crowded and starving, they conserve their resources, often not even copying their DNA. How do they get out of this fix if they
THE GENOME SENDS AN SOS
95
cannot make mutations? Bacteria even can survive by eating DNA,33 but once inside the bacteria, the DNA may have something to say, too. Enzymes cut up incoming DNA. This protects bacteria from being taken over by other genomes, particularly viruses. In this way, bacteria can try out new information in bite-sized pieces34 without letting the incoming DNA call the shots. In H. influenzae in our upper respiratory tract, DNA encoding a gene that looks like it affects whether incoming DNA is chopped up contains a string of 40 repeats of AGTC.35 Since amino acids are encoded by groups of three letters, a slip of four letters throws this protein out of frame and may even create a codon that says, “Stop reading this message.” In addition to this slippery section of DNA, another section of the gene has a lot of changes; variation in this latter section may change where the enzyme cuts DNA36 so that different individuals in a population of bacteria cut incoming DNA differently, generating additional diversity. If DNA encoding DNA-cutting enzymes tends to vary, not all the bacteria in the population will cut the incoming DNA in the same place. This can protect any useful information that is encoded at that spot and will increase the diversity of progeny. The DNA strands of starving bacteria may break.37 Bacteria can “repair” breaks in their DNA either with information from other pieces of DNA or with an error-prone polymerase. In repair by recombination, pieces of DNA can be cut from one helix and pasted into another. The helix that contributes information to the broken DNA may be from the original genome, or it may have been swallowed, perhaps from the same species or perhaps, broken into bite-sized pieces, from another genome. It is a reasonable guess that it was under such stress that two pieces of DNA got together to create the carbohydrate-digesting protein described in Chapter 7. While the mismatch detectors generally enforce the requirement that only very similar sequences can recombine, when mismatch repair is damaged, more diverse sequences are able to be patched into the genome.38 Repair of damaged and lost letters is likely to decrease when cells are starving and struggling. If a missing letter is not replaced, and the polymerase stalls, this leftover damage could trigger the SOS response.39 If the polymerase stalls only momentarily, it is likely to slip, or even fall off. If it slips, genes may be duplicated,40 and repeats can expand or contract. Slips in repeats that are stimulated by breaks in DNA are very relevant to the human genome too. For example, the DNA in children whose parents lived in the shadow of Chernobyl’s radiation had slipped significantly more often when being passed on to them from their parents than the DNA of children in England.41
96
DARWIN IN THE GENOME
Whatever happens in the environment, the extent of the mutation experienced by a genome is affected by the activity of the proteins that copy and repair DNA. Genome-wide increases in mutation can be risky. However, looking closely at mutators that are less effective at mismatch repair or that have sloppy polymerases stuck on shows us that distinct proteins can affect the tendency to mutate in different ways. As the activity of these distinct proteins varies, increased mutation is not distributed completely randomly across the genome, and natural selection acts on the progeny. Clearly, in a population of starving bacteria, we can see that the bacteria are not thriving. Most are likely to be dying. It isn’t easy. But the range of options for random mutation is so vast that any focus that has emerged from repeated cycles of pressure and survival could increase the chance that some might survive again, and that with their survival, a oncefortuitous alignment of the tendency to slip and its biological effect also would survive. This alignment could increase variation in genes that encode mismatch repair proteins and pathogen proteins that stick to the surface of our cells, which may prove useful again to the bacteria in the future. When sequences emerge, such as slippery DNA, which enables reversible changes in the genome, the repertoire of the genome’s progeny is enhanced beyond that which is encoded explicitly in the genome itself. Other changes are more permanent, such as those caused by polymerase “errors” or those that increase acceptance of DNA from outside the genome. The realm of “normal” genome behavior seems to include slides back and forth in genes that have the effect of increasing and decreasing certain classes of mutations, allowing some flexibility around an evolved genomewide optimum balance between fidelity and exploration. As certain types of diversity are generated among progeny, these progeny may come to find the environment not stressful at all. With so much time to learn to handle extreme challenges, some genomes evolved layer upon layer of strategies for finding the exits, the ways out of the extreme stress. Others didn’t, their ship went under, and they died.
9 Journeys through Space and Time When opportunity arises, Life shall be waiting. —Rachel Carson
W
e think of bacteria as individual creatures, lonely killers, but they, no less than we, often live in communities, dependent upon one another for the future existence of their species. Like all members of a community, they send signals to one another that affect their behavior, and that they rely upon for their survival.1 For example, bacteria can become more open to swallowing DNA from their surroundings when they receive the signal that they are among other bacteria, especially those of the same species, that might send over some recipes. In a sense, bacteria get new ideas at meetings. But how do bacteria, without eyes or a memory for names, know whether they are alone? They use what are in fact called quorumsensing signals, small molecules that float around them. The intensity of these signals is determined by the number of other bacteria around and how close they are to one another. Quorum sensing involves turning genes on and off depending upon how many neighbors are around.2 Quorum-sensing bacteria produce and release chemical signal molecules, and there are more of these signals around if there are more bacteria releasing them. It’s as if each person returning to your neighborhood after work began tossing paper airplanes with messages until they were piled up a foot deep at your door, and you
97
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
98
DARWIN IN THE GENOME
saw them through the window. When bacteria detect more than a threshold amount of such a signal, they change which genes are turned on. Among genes that are turned on by a quorum are those needed for bacteria to interact with others, genes involved in virulence, in the ability to send out and to swallow DNA, to make antibiotics, to move, to build filmlike communities, and to form spores. Bacteria can talk to their own species, and to other species. This communication allows bacteria to behave as a community. If there are no potentially helpful neighbors around, a bacterium that is stressed to the max may go into suspended animation, form a spore, and wait for the wind and water to carry it to more favorable circumstances. Of course, forming a spore has its risks. Still, with the organism’s genome more comfortable in its package, no doubt, than my grandparents were probably in steerage, spores are an effective vehicle for travel in space and time. Spores can travel across oceans on the wind, protected from ultraviolet (UV) light by the shade of their own dust clouds.3 Like the great ocean liners, it takes around 5 to 7 days for a spore to travel from Africa to the Americas. African dust can be detected over about a third of the United States; about half of the dust lands on Florida. Given their food preferences, about a quarter of the microbes that might be blown in dust from a drought-ridden field will thrive if their journey takes them to elm trees or crops such as peaches, cotton, and rice. If spores can move that far in space, how far can they move in time? Bacteria can be very patient. Anthrax spores can wait in the soil for over 100 years.4 In an underground cave, in a crystal of salt, bacteria may have been waiting as long as 250 million years—until the scientist evolved who would revive them in a laboratory. The bacteria found deep in the cave are members of the spore-forming genus Bacillus, named by the researchers Bacillus strain 2-9-3. Bacillus are widespread in the soil, in water, and in dust in the air. Some live naturally with us in our intestines, but others, such as Bacillus anthracis (which causes anthrax), can be deadly. The bacteria found in the cave appear to be so closely related5 to a modern species of halophile (salt-loving) bacteria, that some people suspect these bacteria leaked into the salt cave more recently, through channels so tiny that they cannot be seen under the light microscope. Therefore, the age of these bacteria is not settled.6 Still, if other scientists confirm that the bacteria were in their crystal case for 250 million years, they slept through quite a show on the Earth above. While the bacteria slept, protected in brine within a rock, the first dinosaurs appeared. They grew to dominate the Earth, but the bacteria were undisturbed by the sound of Tyrannosaurus rex. The plants on the Earth
JOURNEYS THROUGH SPACE AND TIME
99
first began to create flowers, but through all of this, the bacteria remained in their brine in its rock. An asteroid hit and the large dinosaurs disappeared, and still the bacteria remained within their rock. The tiny mammals diversified and grew. About 50 million years ago, some took to the sea, becoming whales. Then the genome shared by mammals evolved the primate line. The bacteria waited while Homo sapiens appeared, developed agriculture, spread around the Earth, built cities, and invented a microscope. The bacteria remained in their rock in their cave while Homo sapiens built laboratories, learned about DNA, and first began to think about what might be there, within the ancient rocks. Then came a day, between one sunrise and sunset—after over 91,000,000,000 sunrises and sunsets—when a tool brought by a hand reached down, chipped off a piece of the salt crystal, and carried the rock with its tiny passengers to a place beyond the imagination of the tiny reptiles who had once crawled by, near where the rock first formed around the bacteria. The rock was brought from what now is Carlsbad, New Mexico, to a laboratory in Pennsylvania. In this new place, the rock crystal, carefully protected from twentieth-century contamination, was split open. A nutrientrich broth was provided, and the bacteria stirred from their very long sleep and began to copy their genome and to pass it on to their twenty-first-century progeny. We don’t know for certain that they had slept as spores, but we do know that when things begin to get too salty, these bacteria do rush toward becoming spores. A very lucky human might see 36,000 sunrises and sunsets—each of them, if the human took the time to pause and look up, a show-stopper, each one precious. Within the rocks, ancient bacteria are patiently waiting for something, whatever it may be, that is different. It is hard for a biochemist to believe that bacteria really could have waited for a better environment, for as long as 250 million years. After all, DNA depends upon enzyme after enzyme in the cell to protect it; letters lose pieces and fall off, although storage in salt could protect it from some kinds of damage. Stepping into suspended animation as a spore— taking a chance on time, not knowing how long that time might be—seems risky. Over that long a time, I would expect breaks in the DNA and loss of too many of the life-defining code letters. I would think it would be a lethal wait. And yet my biochemist’s certainty that bacteria could not possibly survive the DNA damage accumulated over millennia of millennia is shaken by the example of the hardy bacterium Deinococcus radiodurans. D. radiodurans can live through blasts of gamma radiation 12,000 times as strong
100
DARWIN IN THE GENOME
as would kill people. It can live through a drought so difficult that its DNA crumbles into 100 pieces. D. radiodurans was first noticed growing happily in an irradiated can of meat. Since then it has been found around the world in places where other bacteria cannot thrive, including on freeze-dried rocks in Antarctica and on sterilized medical instruments. It can resist DNA damage from x-rays and UV radiation, and from hydrogen peroxide. If its DNA double helix is severed, broken across both strands in hundreds of places, by 1.75 Mrads in a little over 24 hours, no problem. Most cells of D. radiodurans can patch up their genome without the rearrangements or even increased mutation frequency that we would expect. D. radiodurans has ways of protecting itself against such severe dam7 age. First, it keeps extra copies of its DNA around; growing cells have 4 to 10 copies of their genome. The odds are that with two strands in each helix and 4 to 10 helices, there should be one strand that remains undamaged at each spot. D. radiodurans also is very efficient at repairing double-strand breaks in its DNA by getting information from an intact double helix through a process called recombination. During recombination, an intact strand of DNA can leave its double helix and partner with one strand of the damaged helix. Once there, a section might be cut out of the intact helix and pasted into the damaged one, or its information might be copied into the damaged helix by using the intact strand as a template to attract letters in the right order, replacing the damaged section. We can guess that for each point in the DNA, at least one of the copies in at least one of the bacteria in the colony would have survived the DNAdamaging experience intact, and could be used as a reference. But D. radiodurans does not depend on redundancy alone. In addition to its main chromosome, D. radiodurans has two large extra pieces of DNA that seem to be specialized for survival in adverse conditions. Genes on these two pieces of DNA are turned on during periods of stress. Also, each bacterium is not alone. Not only might it turn to its neighbors to rescue information that was damaged, but it has very efficient systems for using all of the DNA building blocks and carbohydrates washing around it from the cells that don’t survive. D. radiodurans also keeps a full bacterial DNA repair kit, along with, most probably, some things that we do not yet know anything about. Its full genome repair kit includes ways to recognize, cut out, and replace damaged letters, and to repair mismatches. It kicks damaged letters out of the cell before they get patched into the DNA. It is difficult to imagine how the repair kit itself could survive the damaging treatment, but, of course, D. radiodurans brings along extra copies of its repair kit, too. It is hard to believe, but the fact is that these amazing bacteria do survive.
JOURNEYS THROUGH SPACE AND TIME
101
D. radiodurans may be a particular champ at enduring radiation. Still, when even the most ordinary bacteria cast themselves into the future, protected as spores, they too prepare for uncertainty and package DNA repair kits into the spore. This is an evolved strategy for bacteria. The spores of one species of Bacillus was shown to be nearly 100 times more resistant to radiation damage than the bacteria themselves. A bacterium can diversify its genome, or it can wait for its environment to diversify itself. If D. radiodurans can survive through treatments that would seem to mean certain death, perhaps the salt-loving bacteria could indeed have survived for 250 million years in their cold cave. An extraordinary observation has to be repeated in order to be believed, of course, but as scientists we do have to have some humility in the face of data. To help myself imagine how bacteria might survive so long in a cave, I did some calculating that is not accurate, but that gives me some perspective on how to understand this. Here is the calculation that helps me: If everyday bacteria can survive radiation 100 times better as spores than as metabolizing day-to-day bacteria, suppose that means that they can survive 100 times longer than I can imagine a suspended bacterium might survive. If 100 times better really does translate into 100 times longer, then this repair kit carried by the spore, turning a day in the lab to the distances of time, would make 250 million years seem like 2.5 million years. This is still too long a time, but it brings them closer to us (we were already walking upright by then). The lab study in which spores were so much more hardy than bacteria used UV radiation, not drying in salt, and we have to figure out what might protect the proteins in the repair kit itself. There are proteins, called chaperones, that help to stabilize the structure of other proteins, and that are found in spores.8 Do they help? All these musings suggest that if there are a few more biochemical repair tricks in the many undeciphered proteins in bacterial genomes, one or two of the bacteria may be able to crawl to the finish line after all. Just a small number of viable bacteria, out of the trillions upon trillions that may once have been present, have to survive to carry the genome, and with it the species, forward in time. And the survivors do not have to be museum pieces. They can have some genetic changes; they just have to be able to copy their DNA and to divide, to set out on the road of selection in their new environment. So if bacteria that were harvested indeed had survived 250 million years, they may well have had some tricks, too, things that we have not yet learned about by studying bacteria in their comfortable day-to-day lives in our laboratory broths. Bacteria are not the only organisms that package themselves in capsules and trust to the wind, the water, and time. Little animals, like worms, have
102
DARWIN IN THE GENOME
spores. Worms, fungi, seeds, bacteria, and even some crustaceans can survive desiccation. They seem to form a sort of glass around themselves by stringing together certain sugars or other linkable molecular blocks.9 Plants package themselves for the future too, of course,10 waiting within their seeds. A sacred lotus seed brought from China was radiocarbon-dated as 1450 years old; it flowered when it was planted in California. Seeds wait patiently on the forest floor until a forest fire sweeps through. They wait until a large tree lives out its life and falls, creating an opening for light. At that moment, their shoots spring up into the sunshine of the future.
10 Strategies as Targets, Round One: The Pathogens There’s no better way to locate the soft underbelly of a pathogen than through its genome. —Elizabeth Pennisi1
E
volution is not just for the “good guys,” for breeding corn from an ancient grass, for the ones we would pick for our team. Evolution also happens to the species that we might propose for an expendable species list2—our challengers, the genomes that harm us. Through strategic mutation followed by selection, each of our individual immune systems evolves an impressive repertoire of antibodies during our lifetimes to combat genomes that threaten to damage us. But the genomes that attack us also can evolve. Unfortunately for us, when natural selection acts on tumor and pathogen genomes, it favors those that grow well within us. Bacterial genes change—they slide by decreased mismatch repair; they move patches in and out. Bacteria also pass genes around to one another. Tumors evolve and become more aggressive and harder to deal with. Our ability to combat the spread of antibiotic resistance, the emergence of new pathogens, and the growth, spread, and indeed evolution of each tumor all are affected by strategies for evolution that emerge in genomes that threaten us. As entire genomes are sequenced, and as our blindfolds are thus removed, we can begin to comprehend the strategies of our challengers; we
103
Copyright 2003 Lynn Helena Caporale. Click Here for Terms of Use.
104
DARWIN IN THE GENOME
can fight back consciously. We can, of course, search3 for those parts of surface proteins that are essential for infection, those that bacteria can’t change without losing their ability to infect us, and target them with vaccines. With the complete sequences of genomes to examine, with computers to calculate and laboratories to test our ideas, we also can set to work to decode strategies. We are not stopping with the genome either; labs are hard at work trying to find all of the proteins, how they are modified, and where in the cell they are located; this information may be implicit in the genome, but it can be made explicit more quickly by looking at the proteins themselves. Rather than chase after bacteria antibiotic by antibiotic, can we stop the passing around of genes that encode antibiotic resistance without destroying the bacterial information sharing that keeps our planet alive? Can we stop the variation in the coats of bacteria and viruses and parasites so that they cannot run away from our immune systems and so turn our vaccines into yesterday’s news? Can we stop the evolution of tumors within us, stop the selection of tumors that grow better—the ones that spread and evade our immune system, the ones that ignore signals from within our bodies, and even from within the cells themselves, that sense their unregulated behavior and tell them to die? The HIV genome uses a very high mutation rate to keep eluding our immune system. With as many as 10 billion new progeny each day swarming through its unwilling host, HIV can be creative; it can take risks, make millions or even billions of fatal errors, and still have many progeny survive. Fast genetic changes help HIV race to outstrip our hard work in the laboratory, where we in turn race to create new anti-HIV drugs. HIV has evolved a mutation rate that is fast enough to stay ahead of the immune system, but slow enough to avoid a molecular amnesia of what it is and how it grows. If we can change HIV’s mutation rate, we may throw it off its game. To find drugs that kill HIV, we look for things that block its ability to stick to our cells, or we figure out, by knocking them out in the laboratory, which enzymes it cannot live without.4 Once we know its weak spots, we take its essential enzymes—reverse transcriptase, protease, and integrase— and our chemists work overtime making molecules in order to find one that will block each of these enzymes. But if we use the resulting drugs one by one, resistance develops. We must use mixtures of drugs, for the odds are that an individual virus will not be so lucky as to mutate and evolve resistance to all three drugs simultaneously and thus survive. A drug that decreases the high HIV mutation rate—in effect, tying HIV’s genetic hands—should decrease the development of resistance and improve the immune system’s chances of combating HIV, and apparently
STRATEGIES AS TARGETS, ROUND ONE: THE PATHOGENS
105
it does.5 The drug ddC blocks the ability of HIV’s reverse transcriptase to copy its own genome. If HIV has to live in a person who is taking ddC, it has to change its reverse transcriptase so that it is not blocked by ddC. The changes that HIV has to make in its reverse transcriptase in order to avoid ddC also happen to increase the fidelity with which the reverse transcriptase copies its genome. The more carefully HIV copies its genome (the more stable it is), the less it is a moving target. But the difference in fidelity caused by ddC is not enough to make it a cure for AIDS. We can try to search for chemicals that block any possible active mutant of the enzyme. We also can look for chemicals that increase the fidelity of reverse transcriptase dramatically. HIV is the latest in a long line of pathogens that have devastated the human family. New ones emerge from time to time, so we must be vigilant, to spot new threats, to head off what Laurie Garrett calls the coming plague.6 In his book The Forgotten Plague, in a chapter entitled “The Reign of Terror,” Frank Ryan reintroduces us to “one of the most dangerous epidemics in human history,” tuberculosis.7 We know that at least one person who died as long ago as 4000 B.C. was infected with TB. In the nineteenth and twentieth centuries, in some places TB killed one in three adults, a total of about a billion people in those two centuries alone; its impenetrable reign was not ended until the development of one of the first antibiotics, streptomycin. But we still have something to fear from TB. It remains with us, always opportunistic, always pushing back. In the early 1990s it reemerged in a form that was resistant to all antibiotics, threatening enough so that in 1993, the World Health Organization declared a global emergency. More than 7 million people were estimated to be newly infected with active TB in 1998 8 (many of them with bacteria that were resistant to our current treatments); that year, TB killed 3 million people around the world. TB lives in about 1 billion people, kept in check by their defenses, dormant but not dead.9 For this reason, the TB genome was one of the first chosen for sequencing.10 It has just over 4.4 million letter pairs. How does knowing this string of 4.4 million As, Ts, Gs, and Cs help us? The genes don’t have labels. So far, the guess is that the